New molecular virus detection methods and their clinical value in lower respiratory tract infections in children.

During the past decade, several new respiratory viruses and their subgroups have been discovered. All these new viruses, as well as previously known respiratory viruses, can be detected by sensitive PCR methods, which have become popular in the diagnostic workup of respiratory viral infections. Currently, respiratory viruses can be detected in up to 95% of children with lower respiratory tract illness. On the other hand, virus detection rates in asymptomatic children are also high (up to 68%), as are coinfection rates in symptomatic children (up to 43%) and justified concerns of causality have been raised. Imposing progress has been made in developing multiplex quantitative PCR assays; here, several primer sets are run within a single PCR mixture. These PCR assays give a better understanding of the dominant viral infection, of viral infections that may be incipient and of any waning infections than does a single-target PCR. Multiplex PCR assays are also gaining popularity due to their cost-effectiveness and short throughput time compared to multiple single-target PCRs. Our understanding of the indications of virus PCRs and our ability to interpret the results from a clinical point of view have improved. This paper reviews the progress in PCR assays and discusses their role in the diagnosis of lower respiratory tract infections in children.

Introduction

The diagnostics of respiratory viral infections began in 1933 by the discovery of the influenza A virus. Over the next three decades, several other major respiratory viruses were discovered, including enteroviruses, the adenovirus, the respiratory syncytial virus (RSV), the rhinovirus (HRV), the parainfluenza virus (PIV), and the coronavirus. The detection methods consisted of virus culture, antigen detection and serology. These methods yielded an overall virus detection rate of only less than 40% in children with lower respiratory illnesses (acute wheezing/asthma and pneumonia).2, 3

In the 1990s, a new era began with the development of molecular amplification-based techniques (PCR, polymerase chain reaction). Not only did PCR revolutionize the detection sensitivity of known viruses but it also enabled the discovery of altogether new respiratory viruses in the respiratory tract, including the human metapneumovirus (HMPV), new coronaviruses, the human bocavirus (HBoV), HRV-C group and polyomaviruses. Molecular methods detect currently one or more respiratory viruses in up to 95% of children with bronchiolitis, acute wheezing or asthma, and in up to 72% of children with pneumonia (Table 1 ).5, 6, 7, 8, 9, 10 PCR has achieved wide use in clinical virology, and is displacing the conventional methods with the exception of commercial rapid antigen detection tests for influenza virus and RSV. The high sensitivity of PCR has, on the other hand, shown that viruses in asymptomatic children (up to 40-68% during high prevalence seasons) and virus coinfections - or rather codetections - in symptomatic children are common (up to 43% in acute asthma).5, 11, 12, 13, 14, 15 This has raised concerns not only of causality but also whether upper airway specimens really reflect the conditions in the lower airways with regard to viral replication, since the mere presence of a virus may not be etiologically related to the patient's symptoms. In this paper, we review the latest achievements in the molecular diagnostics of respiratory viral infections and discuss their applicability to lower respiratory tract illnesses (LRTI) in children.

Table 1

Virus etiology of lower respiratory illnesses in children.

Illness	Dominant virus	Virus coinfection rate	Total detection rate
Bronchiolitis	RSV	41%	95%
Acute wheezing/asthma	HRV	43%	95%
Pneumonia <12 months,>12 months	RSV, HMPVHRV, HBoV	33%27%	77%72%

RSV, respiratory syncytial virus; HRV, human rhinovirus; HMPV, human metapneumovirus; HBoV, human bocavirus.5, 6, 7, 8, 9, 10

Single-Target PCR

The main advantage of amplification of viral genetic material by PCR is high analytical, and often also diagnostic, sensitivity (Table 2 ). The first PCRs were single-target endpoint assays which allowed product detection only at the endpoint of the PCR amplification. With RNA viruses, reverse-transcriptase PCR (RT-PCR) needs to be used: reverse transcriptase first converts RNA to complementary DNA (cDNA) after which ordinary PCR may be performed. The final PCR or RT-PCR product must be certified to be correct DNA, e.g. by a post-PCR hybridization with a labelled probe. Increased specificity and sensitivity can be achieved by nested PCR in which two sequential pairs of primers are used instead of one (Table 2). The first or outer pair yields a DNA fragment, as is the case in a standard PCR. Then, in a subsequent PCR, a second pair of primers, called nested or inner primers, binds within the first PCR product (amplicon) to yield a shorter DNA fragment. This makes the PCR more sensitive but also more prone to contamination. Moreover, were a non-specific amplicon created in the first round, it would not be amplified in the second round (i.e. increased specificity). PCRs like these result in either a “positive” or “negative” result, but the result is not quantitative.

Table 2

Advantages and disadvantages of PCR-based assays.

Method	Advantages	Disadvantages
PCR-basedvs.conventional methods	• superior sensitivity and specificity• sample type and handling not as critical; the viruses do not need to be infectious, no need for intact cells• easy to automate with current technology• can be used to both detect and quantify the specific virus• can be used as multiplex, enabling detection of several agents in one tube• product can be genotyped• product can be sequenced, enabling molecular epidemiology• provides an early diagnosis; often positive before serology	• false negatives due to inhibition of the polymerase, or other reaction failures• false positives due to contamination• expensive• require proper facilities, instrumentation and procedures• require purification of nucleic acids before the test• require many positive and negative controls• may be clinically even too sensitive, the result needs cautious interpretation• normally, only the virus that is looked for can be found (≠culture)• cannot detect past infections• or distinguish primary from secondary, or acute from chronic infections (≠serology)
Nested PCRvs.normal end-point PCR	• even more sensitive• does not necessarily need downstream hybridization• more specific than end-point PCR without hybridization	• more prone to contamination and needs further precautions• more expensive• takes twice the time• requires 2 separate machines with separate facilities
Real-time (non-quantitative) PCRvs.end-point PCR	• can be highly automated, user friendly with less hands-on time• is faster, has higher precision, resolution and lower variability• allows real-time measurement• can be designed as a probe format which does not need a downstream method for product identification (such as Southern hybridization)• can be designed to genotype the virus or to find a mutation according to the melting points of the products• is less prone to contamination due to the closed system	• more expensive due to instrumentation and reagents• need a computer and sophisticated software
Real-time quantitative PCRvs.real-time non-quantitative PCR	• analyses virus load, either relative or exact• better evaluation of timing of infection• can be used for therapy monitoring	• more difficult to design• needs a known standard dilution series• quantifying cutoffs maybe higher than detection cutoffs, lowering sensitivity
Multiplex PCRvs.singleplex PCR	• detects several different viruses in the same tube• better cost-effectiveness than multiple single-PCR assays (lower expense of reagents, less laborious and shorter throughput time)• less sample consuming• provides naturally several internal controls (one positive reaction increases the negative predictive value of the others), and is more effective in determining the quality of the template• can be coupled with downstream differentiating methods, such as beads or microarrays, which are multiplex technologies that allow detection of 10-2 million analytes in a single sample	• more difficult to design, may need substantial redesigning and optimization• adding multiple primer and probe sets may compromise the accuracy and sensitivity• may provide more variability• due to limited fluorophores or fluorescent spectra, to distinguish more than five viruses usually an expensive downstream detection system is needed• most bead- or microarray-applying amplicon-detection assays lack quantifying properties

Quantitative PCR

Quantitative PCR (qPCR) is a method that not only detects specific DNA or RNA in a sample, but also quantifies it. By modern techniques this is achieved by automatic and computerized PCR devices (Figure 1 ). One or more fluorescent probes, specific for the target DNA, can be applied into the same reaction mixture together with the target, primers, nucleotides and polymerase; now hybridization occurs at each PCR cycle. The emitted fluorescent signals correlate with the amount of PCR product produced. The PCR device measures fluorescence once every cycle, i.e., data is retrieved online and at real time. These features make real-time qPCR more suitable for respiratory virus diagnosis than conventional end-point PCR.

Figure 1

Overview of the real-time qPCR protocol. 1. Before PCR, the genetic material needs to be purified from the sample in a sample-preparation room. 2. In a DNA-free reaction-preparation room, water, buffer, nucleotides, primers, probe and a heat-stabile polymerase enzyme are mixed and aliquoted in reaction tubes, which are taken to the sample room. 3. Purified sample DNA or RNA, standard dilutions, and controls are added to separate reaction tubes, which are then applied to the PCR instrument in a separate amplification room. 4. During PCR the carefully designed specific oligonucleotide primers (in red) are annealed to their complementary regions of the two heat-denatured single-stranded DNA molecules. Progeny strands are synthesized by extension of the primers along the full template strand by the action of the DNA polymerase. This is repeated in 30-45 temperature cycles causing an exponential amplification of the target DNA sequence. At each cycle, the target-specific fluorescent probe (in blue) is hybridized to the increasing amount of targets during which fluorescence is measured. 5. This increase in fluorescence is shown in real time on the screen. The fluorescence curves of the samples are analyzed by the computer and compared with those of the known standard, and the quantity of DNA is calculated. NPA, nasopharyngeal aspirate sample; NTP, deoxyribonucleoside triphosphates (dNTP: dATP, dGTP, dCTP, dTTP); qPCR, quantitative polymerase chain reaction.

Multiplex PCR

Respiratory infections can be caused at least by any one of the currently known 26 respiratory viruses. It was soon realized that PCR was too laborious, expensive, and sample consuming, if a sufficient number of virus analyses was to be made to be clinically meaningful. Many children with respiratory illness turned also out to have virus coinfections – sometimes no less than five respiratory virus species concomitantly.5, 11, 14 Since the clinical signs and symptoms of respiratory virus infections overlap and are seldom pathognomonic for any specific virus and since the sensitivity of conventional diagnostic methods is low, the need for testing for multiple viruses simultaneously by PCR became apparent.2, 3 Multiplex PCR is defined as running more than one PCR analysis in a single reaction tube. This is accomplished by applying more than one set of primers that produce amplicons of varying sizes and are specific for different, individual viral sequences (Figure 1). A sign of professional primer and probe design is careful avoidance of cross- and mishybridization.

Multiplex PCR approaches are becoming increasingly popular for the detection of respiratory viruses, since lower reagent costs and shorter throughput time favour the cost-effectiveness of multiplex PCR over multiple single PCRs (Table 2).17, 18, 19, 20 The current cost of a respiratory virus 13-plex PCR is 140 euros in our hands and the result is available within the same working day. Rapid results are crucial for clinical decisions. Multiplex PCR may be qualitative or quantitative.

Tagging (xTAGR) and bead-based (xMAPR) multiplex technologies are interesting applications of the technique; they allow for the detection of 50–500 analytes (protein or nucleic acids) in a single sample. After a multiplex PCR reaction, target-specific extension primers are attached to an xTAG universal tag sequence which is hybridized to a complementary anti-tag sequence coupled to a particular xMAP bead set. The hybridized beads are finally read by the Luminex system. The multiplex PCR assays have generally been much more sensitive than conventional viral diagnostics23, 24, 25, 26, 27 with specificity and sensitivity similar to those of single-target PCRs (Table 3 ).25, 26, 27, 28, 29, 30, 31 There are, however, slight virus-specific differences between multiplex PCRs and single-target PCRs. These differences seem to vary, and in some studies, multiplex PCRs have been less sensitive than a single-target PCR for RSV, human metapneumovirus, enteroviruses, PIV or adenovirus.25, 26, 28, 29, 31

Table 3

Recent studies comparing different PCR methods in children with respiratory tract illness.

1st author (year)	n	Age	Sample	Methods	No. of viruses tested	Main results
Roh (2008)24	50	84% <6 years	92% NPA8% other	multiplex PCRvirus culture, antigen detection	8	80% of samples showed concordant results
Kim (2009)23	101	mean 7 years	various	multiplex PCRvirus culture	12	Concordance 89%
Raymond (2009)28	221	children	NPA	multiplex PCRmultiplex PCR with primer extension and microarray hybridization	23	Concordance 94%
Gadsby (2010)25	286	median 5 years	respiratory	multiplex xTAG fast assaysingle and multiplex PCRsvirus culture, antigen detection	15	xTAG had sensitivity of 79% and specificity of 100% compared to qPCR
Arens (2010)26	410	paediatric	mostly NP swabs	multiplex PCRvirus culture, antigen detection	6	Sensitivities of multiplex PCR for individual viruses 94-100% and specificities 99-100%
Rand (2011)27	200	61% <18 years	71% upper respiratoty	multiplex PCRmultiplex xTAGvirus culture, antigen detection	8	Sensitivities 86-100% and specificity 100% for both
Jansen (2011)29	133	median 12 months	NPA	single-target PCRmultiplex qPCR	14	Good concordance
Pabbaraju (2011)30	334	43% <2 years	73% NP swabs/NPA	multiplex xTAG classic assaymultiplex xTAG fast assay	12	Overall sensitivities classic 89% and fast assay 78%
Ali (2011)31	225	<12 months	nasal and throat swab	multiplex PCRindividual qPCR	11	Good agreement for most viruses

n, number of samples; NP, nasophryngeal; PCR, polymerase chain reaction; qPCR, quantitative PCR; NPA, nasopharyngeal aspirates.

Upper and Lower Airway Samples

Respiratory viruses are usually identified in samples from the upper respiratory tract. All nasopharyngeal sample types, including nasopharyngeal aspirates (NPA), washes, swabs or brush samples, appear to allow for equally sensitive PCR analysis.32, 33, 34 What has been considered, however, is to what extent upper airway virus findings reflect true viral infections in the lower airways. Paediatric studies using bronchoalveolar lavage or lung biopsy samples have indicated that many respiratory viruses seem to penetrate the distal bronchioles and can be demonstrated at similar detection rates in upper and lower airways.13, 35, 36 Also asymptomatic children have exhibited high (35–43%) virus detection rates in lower airway samples.13, 15, 36

It has been presumed that the most frequent causative agent of the common cold, HRV, occurs mainly in the upper airways, and that it is associated closely with exacerbations of asthma.5, 11 Indeed, HRV has been found by in situ hybridization in lower bronchial biopsy specimens of 45% of young children with recurrent asthma-like symptoms. Findings of HRV also correlated with decreased airway conductance and a history of respiratory symptoms. Another study on tracheal and nasal secretions from paediatric patients with a tracheostomy demonstrated HRV in the lower airways of patients with community-acquired colds; the possibility of nasal contamination was excluded. HRV was detected by PCR both in nasal and tracheal secretions. In agreement with this observation, HRV replicates well at lower airway temperatures in vitro.

PCR studies of induced sputum samples have shown that community-acquired pneumonia is associated with viral infection in up to 72% and viral or bacterial infection in up to 97% of the children affected. Although sputum induction did increase the sensitivity and specificity of diagnosis of lower airway infections, most of the pathogenic viruses were often identified also in samples from the upper airways. Although sputum induction does provide good quality samples from the lower airways, it is not recommended for routine use in children with LRTI since it is unpleasant and does not usually provide unique or additional virological information.

Illness Severity

Some studies have been performed to address the association between virus PCR positivity in upper airway samples and symptom severity. In two studies, HRV was more prevalent in children with respiratory symptoms (mainly wheezing) when adjusted for findings in asymptomatic controls.39, 40 A study involving infants with recurrent respiratory infections showed that the total virus detection rates, the virus coinfection rates and HRV findings did correlate with symptom severity. Studies on young children have shown that RSV is most closely associated with severe bronchiolitis or LRTI.10, 41, 42 Of the HRV strains, the HRV-A and -C groups have been linked to more severe illness than the B group - especially in children.43, 44 HRV-C may cause viremia more often than HRV-A and HRV-B. The virus (including HRV) detection rate in samples from the lower airways of asthmatic children during asymptomatic periods has been the same as of healthy children. The associations between illness severity and virus load or coinfection are discussed in the next sections.

Virus Load

Can qPCR improve the diagnosis of viral infections over end-point PCR? A recent study showed that children with LRTI have an increased total virus load and harbor more viruses in their NPA than children with no LRTI but with upper respiratory tract infection, fever or cough. Individual loads of HRV and PIV-2 were higher in LRTI than in non-LRTI, and qPCR yielded more viruses (including coinfections, where a “dominant virus” was typically identified) than the conventional methods, virus culture and direct fluorescence assay. Similarly, in immunocompetent paediatric patients HRV infection was most closely associated with LRTI in the absence of other viral agents when the viral load was >106 RNA copies/mL. Two previous reports could not identify a link between HRV load and respiratory symptoms in children.47, 48 HRV probably causes asymptomatic infections more often than is generally assumed.

RSV infection is short-lived and the causative agent occurs rarely in asymptomatic controls; hence, a positive RSV result is almost always of clinical relevance, regardless of the quantity of the virus. However, a recent study reported a link between RSV load (but not between HBoV or HMPV loads) and the risk of Streptococcus pneumoniae or Haemophilus influenzae acute otitis media in young children. In another paediatric study, the HMPV load was associated with fever, bronchodilator use, chest radiography, and duration of hospital stay, i.e. obvious markers of disease severity. Studies on the H1N1 influenza A pandemic virus have shown that asymptomatic virus infections occur.51, 52 There is also a direct association between, on the one hand, pharyngeal viral replication and host cytokine responses and, on the other hand, severity of respiratory illness. The virus load decreases markedly in response to antiviral treatment.53, 54

Studies on HBoV1 have consistently shown persistence or recurrence of HBoV1 DNA in the upper airways at low load long after infection. On the other hand, a high HBoV1 load is reportedly associated with acute HBoV infection and LRTI.55, 56 Due to virus persistence, a diagnosis of acute HBoV1 infection should be primarily based on serology or serum PCR, and secondarily on qPCR with a cutoff of >104 genomes/ml of NPA. It is currently not known if low HBoV loads are due to persistence of HBoV DNA, to mucosal contamination, to re-infection or reactivation of latent HBoV, or to prolonged low-level replication and shedding of HBoV, which could be related to damage of the epithelium of the airways by other viral infections or airway disease. Adenoviruses also are frequently found in respiratory coinfections at low viral loads: apparently due to persistence or latency.57, 58 In contrast to HBoV, the adenovirus load correlates poorly with illness severity.55, 59

Looking at virus loads longitudinally (assessed at hospital admission, discharge and post-discharge follow-up visits) in the NPAs of young infants of the general population, a sustained decrease in the RSV load takes place. Another study on children hospitalized for LRTI showed that the viral loads in the NPA decreased in samples taken 3-4 days apart. This was true for most single and viral coinfections, but not for HRV and HBoV infections. There was a marked decrease in the occurrence of positive HRV cultures among adolescents with common cold from day 1 to day 7; most of the patients remained positive by RT-PCR.61, 62 Ultimately, a rise and fall in the number of virus copies during an acute respiratory illness would be needed to demonstrate a causal relationship between virus PCR positivity and respiratory illness.

Coinfections

The distinction between coinfection and codetection has been ignored in most studies, whereby we use the word coinfection in this connection. In a genuine coinfection, two viruses infect the patient at about the same time. In codetection, one or both of the viruses detected may be remnants of a past infection. By combining qPCR with serology, coinfection can be separated from codetection.

Virus-virus coinfections occur in up to 43% of children with LRTI. Such coinfections may contribute to illness severity.5, 7, 10 Interestingly, some viruses occur more often in coinfections than others, e.g. HBoV1 has been detected with other viruses in up to 78% of cases. Fever, leucocytosis, and the use of antibiotics are more common among patients with RSV coinfection than among patients with an infection caused by RSV alone. HRV coinfections have been associated with illness severity in infants with recurrent respiratory infections. In children with pneumonia, viral coinfections have been associated with illness severity,6, 64 but the severity of bronchiolitis was not associated with virus coinfection in one study.

Viral–bacterial coinfection occurs in up to 66% of children with community-acquired pneumonia, and with any combination of codetections between virus and bacteria in up to 84% of the cases.6, 8, 9, 65 The most frequent combinations are Streptococcus pneumoniae with influenza A or HRV. Virus–bacteria coinfections may cause a more severe illness than viral or bacterial infections caused by one pathogenic organism.66, 67, 68 Many reports have shown that mixed influenza (A or B) virus–Staphylococcus aureus infections may cause fatal pneumonia in children.67, 68 Treatment failures appear to be linked to mixed viral–bacterial pneumonias in children.9, 69 In children with invasive pneumococcal disease, viral coinfections are common (34%) and are probably associated with higher mortality. In adults, there is also a link between the severity of asthma exacerbations (which are usually associated with viral infection) and the prevalence of invasive pneumococcal disease.71, 72

Clinical Interpretation of PCR

Can we rely on PCR-based respiratory virus detection alone? This question was asked often in the early days of PCR. Profound methodological improvements over the years have reduced the likelihoods of false positive and false negative results. The measures to improve the performance of PCR assays include stringent precautions to avoid contaminations, e.g. using filtered tips, separate rooms and laminar flow hoods, high-standard nucleic acid preparations kits, internal positive controls and numerous negative controls. Inadequate reporting of experimental details regarding qPCR can still impede our ability to evaluate critically the results of scientific papers and the reliability of the specific reported qPCR method. Helpful and useful guidelines are available aiming at improving the quality of reporting of new qPCR methods, e.g. the Miminum Information for publication of Quantitative real-time PCR Experiments (MIQE).

Clinically, the detection rates have generally been higher in symptomatic than asymptomatic individuals. The detection rates have also been higher in asymptomatic individuals with a chronic condition than in asymptomatic individuals without such a condition; this argues against a significant role for false positive results. Perhaps a still more important technical matter is the sensitivity of the PCR – it is generally higher than the sensitivity of conventional methods,23, 24, 25, 26, 27 and therefore, PCR is likely to detect more viruses which may occur at different phases of infections, e.g. incipient, acute and past infections. The risk of false negative findings increases as the number of primer and probe sets is increased, but this risk is ameliorated by inserting internal controls in the reaction tube in order to reveal inhibiting factors in the sample.18, 19, 20

Several findings suggest that PCR is likely to detect true respiratory infections irrespective of symptoms. First, although many respiratory viruses do infect the lower airways,13, 36 their persistence in respiratory samples in general is of a rather short duration – with the exception of HBoV1.14, 48, 56, 74, 75 Second, despite high virus-detection rates in both lower and upper airway specimens of asymptomatic children,13, 15, 36 symptom severity is associated with single-virus detections, multiple-virus detections and virus loads, as discussed above.14, 39, 40, 41, 42 Third, PCR findings of HRV and RSV in nasopharyngeal aspirates correlate with the appropriate systemic immune responses in young wheezing children.76, 77, 78 These data argue against the claim that PCR findings in asymptomatic subjects would be due to residual nucleic acids from preceding respiratory infections.

In a symptomatic subject, a positive virus PCR result may identify the causative agent of the current illness (e.g. RSV), a causative agent of an ongoing asymptomatic infection (e.g. HRV), prolonged presence of virus from a past infection (e.g. HBoV), a latent infection or colonization (e.g. adenovirus), or reactivation (e.g. Epstein-Barr virus and human herpesvirus 6). In an asymptomatic subject, a positive virus PCR result, if not due to laboratory contamination, may be due to 1) symptoms and signs that are not recognized, 2) truly asymptomatic acute infection,52, 79 3) asymptomatic persistence, or 4) mucosal contamination.

Clinical Applicability of PCR

All respiratory viruses can be detected by sensitive PCR-based methods. Since acute symptomatic respiratory illness is often accompanied by multiple virus detections in one sample, it may be difficult to prove, against the Koch postulates, that the mere presence of a pathogen would imply that it is the causative agent of the current illness.5, 14 There is a clear need to move towards quantitative multiplex diagnostics.21, 80, 81 qPCR allows better evaluation of the time course of the infection and of therapy monitoring than end-point PCR. The prevalence of causative viruses exhibits not only seasonal variation, but also considerable year-to-year variation (virus epidemics). The results of a large Swedish study (7853 samples obtained during 36 consecutive months) arrived at a recommendation of using a similar test panel all year round. The routine test panel may, however, need to be redesigned or supplemented by qPCRs, depending on the specific epidemiologic situations to increase sensitivity and enable quantification. Overall, it seems that a high viral load is linked with illness severity and LRTI. Quantitative multiplex PCR may provide an estimate of which viruses are active at the moment and which are emerging or on their retreat. Upper respiratory samples seem to be adequate for clinical routine. Matters that still need elucidation are defining cutoff levels for acute virus infections and standardization of sampling, i.e. how to take comparable respiratory samples over time and interindividually for qPCR.63, 82 When used together, qPCR and serologic assays provide a more complete picture of respiratory viral infections, especially HBoV1.

Excluding influenza viruses, there is yet no specific treatment or vaccine available for treating or preventing respiratory virus infections. Viral detection may, however, have practical importance in isolating infected patients in hospitals or in long-term care settings to prevent spread of the contagious disease. The diagnosis of a viral infection may also be important in the battle against unnecessary antimicrobial treatments, although results from a recent report based on 12-36 hour response time in virus detection had no clinical impact in this respect. Bacterial and viral coinfections may enhance the severity of the illness synergistically by 1) destruction of the respiratory epithelium by the viral infection, which may facilitate bacterial adhesion, 2) virus-induced immunosuppression that may cause bacterial superinfections, and 3) inflammatory response to viral infection, which may upregulate the expression of molecules that bacteria utilize as receptors. The full extent of the clinical significance of mixed viral–bacterial super- or coinfections remains undecided.

Interestingly, data are accumulating to suggest that the susceptibility to certain viral infections could be a marker of chronic pulmonary inflammatory processes. Paediatric studies have shown a strong link between susceptibility to HRV-induced early wheezing and the development of recurrent wheezing or asthma later in life (Table 4 ).87, 88, 89, 90, 91, 92 There is some evidence that young, wheezing children prone to HRV infections will respond to systemic glucocorticosteroid treatment at a high likelihood.89, 93, 94 HRV infections have been linked to an atopic diathesis in young wheezing children,11, 93, 95, 96, 97, 98 which partly explains the observation of the effect of systemic glucocorticosteroid treatment.

Table 4

Human rhinovirus detected in upper airway samples of young children with acute wheezing; the risk of recurrent wheezing within 1-2 years and the risk of asthma at age 5-7 years.

1st author (year)	Setting	Age (months)	Index group	Comparator group	Risk of recurrent wheezing within 1-2 years OR or HR (95% CI) adjustment	Risk of asthma at age 5-7 years OR (95% CI) adjustment
Lemanske (2005)88	Outpatients, birth cohort at atopy risk	<12	HRV+ wh+, n=43	HRV+ wh-, n=119	OR 10 (4.1, 26)multiple	-
Lehtinen (2007)89	Inpatients, non-selected population	3-35(1st episose)	HRV+wh+a, n=37	RSV+ wh+ b, n=43	HR 5.05 (1.00-25.41)Age, atopy, maternal asthma	-
Midulla (2011)92	Inpatients, non-selected population	<12(1st episose)	HRV+ bronchiolitisn=37	acute non-respiratory diseasen=39	OR 3.3 (1.0-11.1) multiple	-
Kotaniemi-Syrjänen (2003)87	Inpatients, non-selected population	<24	HRV+ wh+, n=20c	HRV-, wh+, n=43	-	4.1 (1.0 - 17)yesd
Kusel (2007)90	Outpatients, birth cohort at atopy risk	<12	HRV+ wh+, n=34	LRI+ wh-, n=193	-	3.2 (1.1 - 9.5)atopy
Jackson (2008)91	Outpatients, birth cohort at atopy risk	<12	HRV+ wh+, n=45	HRV+ wh-, n=214	-	2.7 (1.4 - 5.3)yese
Jackson (2008)91	Outpatients, birth cohort at atopy risk	12-23	HRV+ wh+, n=37	HRV+ wh-, n=222	-	6.5 (3.1 - 13.7)no

OR, odds ratio; HR, hazard ratio; CI, confidence interval; HRV, human rhinovirus; wh, wheezing; RSV, respiratory syncytial virus; LRI, lower respiratory tract illness.

HRV diagnosed alone or together with any other viruses.

RSV diagnosed alone or together with any other viruses except HRV.

Single infections.

Independent of age, sex, and atopic dermatitis.

Independent of aeroallergen sensitization.

Conclusions

All respiratory viruses can be detected with modern, highly sensitive and specific PCR methods, which are likely to detect the causative agent of respiratory infections of patients with or without symptoms. Clinically, multiplex quantitative PCR is currently the most attractive option since it is cost-effective and the result can be read within hours after sample collection. Because some sensitivity is inevitably lost when multiple primer sequences are included, primers may need redesigning, or multiplex PCR may need to be supplemented by singleplex PCRs, depending on the epidemiological or clinical context. Since multiple viral infections are not uncommon, qPCR may give information on the most active, incipient or weaning infections. Combining qPCR and serologic assays provides a more complete picture of respiratory viral infections.

The hurdles of this approach relate to standarization of sampling and to interpretation of multiple virus codetections. Considering that practically all viruses can reach the lower airways and are typically also present in the upper airways, upper airway sampling can be considered adequate for the viral diagnosis of lower airway illnesses. Usually a high virus load and multiple virus detections correlate with lower-airway involvement. Although lower airway sampling, typically bronchoalveolar lavage or induced sputum, increases the sensitivity and specificity of diagnosing lower airway viral infection, these invasive and uncomfortable sampling methods should be reserved for special situations, e.g. for investigating immunocompromized patients in an intensive care unit setting.

Some recent findings suggest that certain viral infections could be markers of pulmonary inflammatory processes. This is especially the case for HRV-induced wheezing which is a high risk factor of childhood asthma. In the future, identification of causative viruses of all hospitalized children with early asthma-like symptoms may be possible and a goal worth pursuing.

Educational Aims

To give an overview of the modern respiratory virus PCR assays

To review the applicability of these methods to lower respiratory tract illnesses in children

To help physicians to use and interpret respiratory virus PCR assays optimally in their clinical practice