The role of infections and coinfections with newly identified and emerging respiratory viruses in children

Metapneumovirus was first recognized in 2001 in the Netherlands from nasopharyngeal aspirates collected during a 20-year period in 28 hospitalized children and infants with acute respiratory tract infection (RTI) having signs and symptoms similar to that of RSV infection [4]. The virus genomic sequence was identified by using a randomly primed PCR protocol and revealed to be closely related to the avian pneumovirus, a member of the Metapneumovirus genus, in the Paramixoviridae family, Initial studies following the first hMPV identification indicate that it causes upper and lower RTIs in patients of all ages, but mostly in children aged below 5 years [12‐17]. A large epidemiological retrospective study examined nasal washes collected over a 20-year period during acute respiratory illnesses in an outpatient cohort of children [18]. Over the entire study period, hMPV was detected in 1%-5% of pediatric upper RTIs (UTRIs), with variation from year to year. Several reports indicate that hMPV is a commonly identified cause of pediatric lower RTIs, and is second only to RSV as cause of bronchiolitis in early childhood [16, 19‐25]. While bronchiolitis, is the most common presentation of hMPV illness, other reported syndromes have included asthma exacerbation, otitis media, flulike illness, and community-acquired pneumonia [12, 15, 26‐29]. Several studies have found hMPV –RSV co-infection rates of approximately 5-14% [30‐34]. Nevertheless, in a study conducted in the Netherlands in children admitted to hospital for lower RTIs (LRTIs), no virus co-infection between RSV and hMPV was detected [35]. Different controversial reports suggest an association between RSV-hMPV coinfection and an increase in the disease severity or the absence of an association between dual infection and disease severity. Greensill and colleagues [36] reported a 70% rate of co-infection with hMPV in a cohort of infants with critical RSV bronchiolitis who required intensive care in the United Kingdom, suggesting that dual infection with RSV and hMPV may predispose for a more severe disease. In another study from the United Kingdom, hMPV and RSV co-infection was associated with increased disease severity and higher risk of admission to the pediatric intensive care unit [37]. Similar findings are supported by other studies suggesting that in young children, coinfections with RSV and hMPV are more severe than infections with either RSV or hMPV alone, requiring a longer hospitalization and supplemental oxygenation [38, 39]. However, such synergistic association has not been found in other population-based and case–control studies of hospitalized children [16, 33‐35, 40‐42]. In particular, two studies evaluated the epidemiology of hMPV coinfection in children with LRTI caused by RSV and demonstrated no hMPV and RSV co-infection in mechanically ventilated children suggesting that co-infection with hMPV is not associated with a more severe course of RSV-LRTI [35, 42]. In addition, in a prospective 2-year study in hospitalized infants with acute respiratory diseases, the role of RSV as a major respiratory pathogen was not influenced by the co-circulation of other emerging viral agents with similar seasonal distribution. In particular, RSV-hMPVs co-infections were significantly observed in less severe respiratory disease when compared to unique RSV infections [34]. The possible synergistic interaction between hMPV and the severe acute respiratory syndrome (SARS) coronavirus was also suggested during the 2003 SARS outbreak in Hong Kong and Canada [43, 44]. In one case report, in an infant with SARS CoV infection, fatal encephalitis was correlated with hMPV infection as hMPV RNA was detected post-mortem in brain and lung tissue [26]. Nevertheless, in experimental studies performed in macaques, a synergy between hMPV and SARS was not confirmed [45]. In addition, infections of hMPV with respiratory viruses different from RSV, have also been occasionally reported but no sufficient data are available to discuss epidemiology or association with clinical disease presentation [34, 46, 47] (Table 1).

Following the discovery of SARS-CoV, other human coronaviruses, HCoV-NL63 and HCoV-HKU1, were identified and recognized to be common causes of community-acquired respiratory infections.

HCoV-NL63, a member of the group I coronaviruses, was first detected in 2004 in the Netherlands from a child with bronchiolitis by using a new method for virus discovery based on the cDNA-amplified restriction fragment−length polymorphism technique (cDNA-AFLP) [3].

HCoV-HKU1, a group II coronavirus, was first detected in Hong Kong in 2005 from an adult patient with chronic pulmonary disease [1]. All attempts to grow a virus from his respiratory secretions failed until recently [48], but coronavirus RNA was initially detected by RT-PCR using pol gene consensus primers.

Like other coronaviruses, NL63 and HKU1 can also be detected in individuals of all ages, including elderly patients with fatal outcome [49] and those with underlying diseases of the respiratory tract [1]. However most frequently, the newly discovered coronaviruses are reported in 7 to 12-month old children with both upper and lower RTIs [34, 49‐53]. In studies conducted in children hospitalized with RTIs in China, from 2.6% to 3.8% of patients were positive for HCoV-NL63 and, in addition to causing upper respiratory disease, HCoV-NL63 was found in croup, asthma exacerbation, febrile seizures, wheezing and high fever cases [54, 55]. The occurrence of co-infection with NL63 and other respiratory viruses, including other human coronaviruses, RSV, PIV, influenza A and B viruses and hMPV has been reported [34, 55‐58]. In a large study from Germany evaluating children under 3 years of age with LRTIs, most co-infections were with RSV-A, probably because of the high percentage of RSV-A infections and an overlap in seasonality. In addition, double infection of NL63 with RSV-B, and with PIV3 occurred in a minority of cases. HCoV-NL63 co-infection with RSV-A occurred predominantly in the hospitalised patients in contrast to HCoV-NL63 co-infections with PIV3 that were exclusively present in the outpatient group [59].

Following the first identification, HKU1 was found in respiratory samples from elderly patients and children mainly with underlying diseases [1, 33, 52, 60]. The most common symptoms are rhinorrhea, fever, and abdominal breath sounds [33], but pneumonia, bronchopneumonia, bronchiolitis, and acute asthma exacerbations were also described in children in China [61, 62].

In a study aimed to evaluate the overall prevalence of 10 respiratory viruses in children with acute LRTIs in China from 2006 to 2009, 73.47% of the HCoV-HKU1 and HCoV-NL63-positive samples tested positive for at least one other virus, most commonly HRV and RSV [54]. Similar data describing a high rate of coinfection of coronaviruses with RSV has also been previously reported [63]. In a report from the UK both dual and single infections associated with respiratory outcomes were observed for HKU1 as well as for NL63 and OC43 coronaviruses [51]. In this study a high number of coinfections was observed for HKU1, NL63 as well as for OC43, mostly with RSV. Similar rates of lower and upper infections were observed in single HKU1 or OC43 infection compared with coinfection, whereas both URTI and LRTI were observed more frequently in single compared to mixed infection with NL63. No differences in clinical outcome were observed between single and dual infections with RSV and Coronaviruses NL63, HKU1 or OC43 indicating that RSV may presumably facilitate coronavirus infection without increasing disease severity. However, in the same study considering viral load data, a role of these coronaviruses in coinfections in respiratory disease was suggested. In fact no differences were observed when coronavirus load was evaluated in single infection and in RSV coinfection, indicating both that infection with another respiratory virus does not affect the ability of NL63, HKU1 or OC43 to establish infection and replicate, and that detection of coronaviruses in mixed infection should not be considered a secondary infection without contribution to disease pathogenesis [51]. This quantitative evaluation is in contrast with previous results obtained by van der Hoek and colleagues describing a significantly lower HCoV-NL63 viral load in patients coinfected with RSV or PIV3 than in patients infected with HCoV-NL63 alone [59]. However, the prolonged persistence of HCoV-NL63 at low levels, the different time of sampling relative to the time of disease onset, or the use of different diagnostic technologies could have affected these evaluations (Table 1).

Human Bocavirus (HBoV) was discovered in 2005, in Sweden by Allander and colleagues by using a large-scale molecular viral screening technique including DNase sequence-independent single-primer amplification [6]. Since initial observations, several studies have reported the prevalence of human Bocavirus infection all over the world ranging from 2 to 21.5%, mainly in children younger than 3 years of age where it has been associated with upper and with lower RTIs [64‐69]. In a study from Norway, HBoV was detected in 12% of children with RTI and it was the fourth most common virus after RSV, HRV and hMPV [70]. Recently, in children with radiographically confirmed community acquired pneumonia in which 17 respiratory viruses were tested during the acute phase of the disease, HBoV was the most frequently detected virus after RSV and HRV [71]. Since the discovery of the first HBoV (HBoV1), three other related bocaviruses (HBoV2, 3 and 4) have been identified in stool samples and associated with gastrointestinal diseases [72, 73]. Serological studies on HBoV1 are in line with molecular data.

Serological studies have shown that the mere presence of HBoV DNA in the respiratory tract is not proof of an acute primary infection [65, 74‐76]. These data are also supported by studies on consecutive respiratory samples showing that HBoV DNA can persist for several months in the respiratory tract [77‐79]. Prolonged viral shedding could explain both data reported in some papers in which HBoV DNA was found more often in asymptomatic than symptomatic cases [79, 80] and the high percentage of co-infections. In fact, HBoV infections tend to be associated with high rates of coinfections with other viral pathogens such as HRV, adenoviruses, RSV, as well as with bacteria such as Streptococcus spp and Mycoplasma pneumoniae[46, 54, 68‐70, 77, 81‐83]. Characteristics of persistence and high frequency of coinfections have led to a debate over its role as a true pathogen [84]. Our current knowledge of HBoV infection suggests that the virus is sometimes a passenger and sometimes a pathogen in acute respiratory tract disease and that diagnosis should not be solely based on qualitative PCR in respiratory samples. Indeed, in many studies a positive correlation was seen between respiratory illness and high copy numbers of HBoV1 DNA or the presence of HBoV1 monoinfection [68, 75, 85‐87]. A study performed by Allander and colleagues suggests that acute HBoV infections are associated with the presence of viral DNA in the blood of patients. In fact, HBoV DNA was reported more frequently in patient blood during the acute symptoms than after recovery [68]. In addition, high load and viremic HBoV infection were associated with respiratory tract symptoms, while detection of a low viral load in the nasopharinx alone resulted to have no clinical relevance [68]. Other studies confirmed that HBoV is the most probable cause of respiratory tract disease if the patient has a single infection and high viral load in NPA and viremia [65, 70]. However, despite these diagnostic challenges it is becoming increasingly evident that HBoV1 is an important respiratory pathogen [88]. Severe and life-threatening disease has been recently well documented in a 8-month-old child with acute respiratory distress attending an emergency department in Germany [89]. Don et colleagues [76] found serological evidence of an acute HBoV infection in 12% of children with pneumonia and in more than half of these cases with single HBoV infection. In most cases a significant rise in IgG antibodies between paired sera was found in children admitted to hospital for radiologically confirmed pneumonia. IgM antibodies were also detected in all but one patient. This study suggests that HBoV may be a fairly common cause of pneumonia in children Table 1.

Two new polyomaviruses were identified in 2007 in respiratory tract samples following large scale molecular screening using high throughput DNA sequencing of random clones [5, 7] and have been named after the institutes where they were found: KI (Karolinska Institute) polyomavirus (KIPyV) and WU (Washington University) polyomavirus (WUPyV). Data on seroprevalence indicate that infection is widespread ranging from 54.1 and 67% for KI and from 66.4% and 89% for WU in North American and German blood donors [90, 91]. Since their first identification, KI and WU viral sequences have been confirmed worldwide in respiratory samples from children with respiratory tract disease ranging from 0.2% to 2.7% and from 1.1 to 7%, respectively [91‐97]. However WUPyV and KIPyV were found at similar frequencies in control groups without respiratory diseases so the link between these polyomaviruses and acute respiratory diseases remains speculative [94, 96, 98].

Careful analysis is complicated by high co-infection rates with other well-characterized viral respiratory pathogens. A co-detection rate of 74% has been observed for KIPyV and rates ranging from 68 to 79% for WUPyV [94, 95, 97]. Therefore, in a recent study in Southern China, hospitalized children with WUPyV infection displayed predominantly cough, moderate fever, and wheezing, but were also diagnosed with pneumonia, bronchiolitis, upper respiratory tract infections and bronchitis. As in most of infected children a single WUPyV infection was detected, it was suggested that the newly described polyomavirus can cause acute respiratory tract infection with atypical symptoms, including severe complications [99]. Nevertheless these data have to be confirmed in further studies.

The presence of WUPyV and KIPyV in samples from children but not from immunocompetent adults suffering from LRTIs suggests that these viruses primarily infect the young population [100]. A correlation between immunosuppression and reactivation of the two novel polyomaviruses has been suggested in immunocompromised patients [101] and in AIDS patients at the molecular level [102], but no evidence of a role of these viruses as opportunistic pathogens has been given.

Overall, these data support the hypothesis that, in analogy with BK and JC polyomaviruses, KIPyV and WUPyV can establish persistent infection, and that virus replication may increase in immunocompromised hosts. However, in a recent study on immunocompetent and immunocompromised adult patients, real-time PCR detected KIPyV and WUPyV in 2.6% and 4.6% of HIV-1–infected patients respectively and in 3.1% and 0.8% of blood donors respectively, while no association was found between CD4+ cell counts in HIV-1 positive patients and infection with KIPyV or WUPyV [103].

KIPyV and WUPyV are also incidentally detected in adults with community acquired pneumonia, in immunocompromised hosts, and in patients with lung cancer; they are more often found in patients suffering an underlying medical condition and coinfections with KIPyV and WUPyV with other respiratory viruses are common [92, 103, 104]. A recent study evaluating the prevalence and viral load of WUPyV and KIPyV in respiratory samples from immunocompromised and immunocompetent children showed that the prevalence of WUPyV and KIPyV is similar in hematology/oncology patients compared with that of the general pediatric population [105]. High co-detection rates with other respiratory viruses, mainly RSV and enterovirus or rhinovirus, were found for WUPyV and KIPyV in both groups, in analogy with previous reports. However, higher viral loads for KIPyV in the immunocompromised group were detected, suggesting that there may be an increased replication of this virus in this population.

A similar association was not observed for WUPyV. Furthermore, in the immunocompromised group, infection with either virus occurred in older children compared with controls, which may indicate viral-reactivation Table 1.

A novel HEV (EV104) genotype was first identified in Switzerland in 2010 from respiratory samples collected during 2004–2007 in 8 children with respiratory signs and symptoms and acute otitis media [121]. In a following epidemiologic study conducted in Italy, five strains of the new EV104 genotype were detected in patients with respiratory diseases [122]. Patients were aged 2 to 62 years and most had underlying diseases; in immunocompetent patients EV104 was associated with chronic rhinopharyngytis, whereas in immunocompromised patients with symptoms of acute respiratory tract infection including wheezing, fever and rhinorrhea. Only in one out of 5 patients a coinfecting virus was detected.

EV109 was first discovered from a case of acute respiratory illness in a Nicaraguan child in September 2010; EV109 isolates were then detected in 1.6% of respiratory samples of children with influenza like illness (ILI) in Managua, and recognized to have a pathogenetic role in the illness [2]. After this characterization, a species C of HEV distantly related to EV109 was retrieved from a rectal swab of a deceased patient during an outbreak of flaccid paralysis in Congo; in this sample poliovirus and other neurologic, enteric and respiratory viral pathogens were not detected [123]. The global distribution of EV 109 is currently unknown and only a few studies have been performed yet to evaluate epidemiological features of this infection. [124, 125] To date the pathogenic role of these new enterovirus strains has still to be defined in larger clinical and epidemiological surveys Table 1.

There are fourteen known HPeV genotypes that were isolated mainly in young children [126‐134]

Two recent studies have investigated the involvement of HPeVs in respiratory diseases reporting a low frequency of detection and a lack of clear disease association [135, 136]. Both studies detected HPeV infections mostly in children below 5 years and the most commonly identified parechoviruses were HPeV1 and HPeV3, while HPeV4, HPeV 5 and HPeV 6 were detected in a minority of cases. However, in addition to a low HPeVs prevalence in respiratory samples, a high rate of co-infection with other respiratory viruses was observed in HPeV positive samples, making it difficult to define a pathogenic role of these new HPeV genotypes in child respiratory infections.

In March 2009, a novel reassortant influenza strain (A/H1N1v) was discovered in Mexico as an infective agent in humans [137, 138]. From April 15 through May 5, 2009, a total of 642 confirmed cases of human infection with the outbreak strain of H1N1v were identified in 41 states in the USA [139, 140]. Of the patients with confirmed infection, 9% required hospitalization. The age of hospitalized patients ranged from 19 months to 51 years and among the hospitalized patients for whom data were available, 18% were children under the age of 5 years. Multiple reports have described the clinical features of infection with this novel virus, which are similar or milder to those of seasonal influenza, [141, 142]. The spectrum of clinical presentation varies from self limiting respiratory tract illness in most cases, to primary viral pneumonia resulting in respiratory failure, acute respiratory distress, multi-organ failure and death, most of them occurring in patients with underlying medical conditions [142].

During the A/H1N1v flu wave, due mainly to the lack of cross-neutralizing anti-influenza antibodies [143, 144] or the presence of co-morbidities, more children and younger adults were infected by the pandemic flu strain and had serious disease than in the seasonal epidemic. For this reason they were identified as a particular risk group, together with pregnant women [145, 146]. Young children presented a higher attack rate than older adults and a greater mortality rate than previously observed with classical seasonal flu [147, 148]. In a study conducted in England from June 2009 to March 2010, a childhood mortality rate of 6 per million population was reported. The rate was highest for children less than 1 year and with severe pre-existing disorders [149]. Similar data are reported in other studies highlighting that most severe cases occurred in children with known comorbidities [150‐153].

The occurrence of co-infection with influenza A H1N1v and other respiratory viruses has been reported in few studies. In a study conducted in France and aimed at investigating respiratory pathogens involved in ILI during the early weeks of the 2009–29010 H1N1v diffusion, 19% of samples positive for H1N1v were also positive for other respiratory pathogens. In mixed infections, HRV was the more frequent co-pathogen being detected in 13.2% of all samples positive for H1N1v [154]. However, HRV infections represented nearly half of non-H1N1v viral infections and the increase of H1N1v positive cases was associated with a rapid decline of HRV infections. In addition, the frequency of virus co-infection was slightly but not significantly higher in samples positive for H1N1v as compared with samples positive for other respiratory pathogens and viral mixed infection both with H1N1v and other viruses was not associated with a different clinical presentation. Similar co-infection data are reported in a prospective study evaluating a combination of two RT-PCR DNA microarray systems in virological routine diagnosis of ILI [155]. In both studies conducted in France, in the same H1N1v early pandemic period, few RSV infections were reported compared with epidemiological data in the same period of the last four years. The delayed and reduced circulation of RSV observed in 2009–10 compared with 2008–09 suggests that the early circulation of the 2009 pandemic influenza A(H1N1) viruses had an impact on the RSV epidemic [156]. A significant inverse relationship between HRV and A H1N1 pandemic virus was also reported, suggesting that the presence of HRV reduces the risk of infection by the H1N1 virus and thus, indirectly, the spread of the virus [156] Table 1.

Torque tenovirus (TTV), is classified in the new genus Anellovirus[157]. TTV produces long-lasting (possibly life-long) viremia in about 80% of apparently healthy individuals of all ages and living in different countries[158]. Although no evidence was obtained that TTV is a direct cause of acute respiratory diseases, it was observed that average viral loads are considerably higher in children with bronchopneumonia than in children with milder illnesses, regardless of the presence of common respiratory viruses [159]. Further studies could not confirm the association, but documented a link between TTV infection in children and asthma, suggesting that TTV might be a contributing factor in the lung impairment caused by this condition [160]. Finally, it was documented that TTV is able to infect respiratory ciliated cells and that these cells are potentially able to support viral replication [161]. Far from being conclusive, the data suggest that these viruses may replicate efficiently in the respiratory tract of children with and without acute respiratory infections by other respiratory viruses, and that although TTV does not have a clear pathogenic role in acute respiratory diseases, it may influence the clinical presentation of the disease.