INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Viral respiratory infection is associated with wheezing illnesses and asthma exacerbations throughout childhood. Wheezing during infancy is defined by seasonal epidemics in the United States of respiratory syncytial virus (RSV) infection (1). In school-age children, rhinovirus is the predominant common cold virus and can be cultured during as many as one-third of wheezing attacks (2). Reverse transcription-polymerase chain reaction (RT-PCR) can enhance the detection of rhinovirus and other RNA viruses, and investigations of children and adults with asthma using RT-PCR suggest that rhinovirus may be present during a larger proportion of exacerbations than previously appreciated by viral culture alone (3, 4). The possibility of improved detection of naturally occurring rhinovirus infection has encouraged new investigations into the role of respiratory viruses in childhood wheezing and a reexamination of the relationship between viral respiratory tract infection and airway inflammation.

In both cross-sectional and prospective studies, IgE antibodies to inhalant allergens are a major risk factor for wheezing in school-age children (1, 5, 6), and it was suggested as early as three decades ago that susceptibility to wheezing with viral respiratory infection could be related to "allergic" factors already in place (7). Subsequent investigations have described elevations of serum total IgE levels after infection with RSV, parainfluenza, and rhinovirus infection (8, 9). In our previous cross-sectional emergency room study, children who had rhinovirus infection as detected by culture in combination with a positive radioallergosorbent testing (RAST) to common aeroallergens had the strongest odds of wheezing (1). Augmented eosinophil responses in the respiratory tract secretions have also been described during viral infections. In this regard, experimental rhinovirus infection prior to allergen challenge has been shown to enhance the late-phase eosinophil influx in bronchoalveolar lavage, as well as airway hyperresponsiveness to histamine (10, 11). Moreover, a persistent eosinophil infiltration has been demonstrated in the bronchial mucosa and submucosa after experimental infection with rhinovirus (12). Elevated levels of eosinophil cationic protein (ECP), a toxic granular protein in eosinophils, have also been described in infants with RSV bronchiolitis, but not in nonwheezing RSV illness (13). We have also observed elevated ECP in the nasal washes of a significant proportion of wheezing infants as well as older children requiring emergency treatment (14). In this latter study population, a comprehensive analysis of viral pathogens, including culture, antigen testing for RSV, and RT-PCR for coronavirus and rhinovirus has been completed. We have examined the results in relation to presence of wheezing, age, atopic status, and eosinophil findings in the nasal washes and blood from study participants.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Populations

Between January 1993 and April 1994, 130 study participants were recruited throughout the school year (excluding June, July, and early August) from pediatric patients treated in the Pediatric Emergency Department at the University of Virginia Health Sciences Center. The children ranged in age from 2 mo to 16 yr. Seventy-one children had wheezing documented by auscultation and were treated with at least a ₂-agonist by inhalation. Patients with a history suggestive of bronchopulmonary dysplasia or who had received systemic steroids within the previous week were excluded. One wheezing patient was excluded after enrollment because of incomplete specimens. Fifty-nine children seen in the emergency department for nonrespiratory illnesses and who were without cold symptoms or a history of wheezing were enrolled as control subjects. Patient demographic information obtained from questionnaires and hospital records has been previously reported (14). There was a history of prior wheezing illnesses reported for 73% (15 of 22) of the wheezing infants younger than 2 yr of age and for 94% (45 of 48) of the older wheezing children. The study was approved by the Human Investigation Committee at the University of Virginia, and informed written consent was obtained from parents.

Specimen Collection and Cell Analyses

Nasal washes were obtained as previously described (14). Briefly, 2 ml of physiologic saline were instilled into each nostril while children were supine. Wash fluid and secretions were immediately aspirated into a sterile 70-ml collection trap (Sherwood Medical, St. Louis, MO) by gentle wall suction through the tip of a flexible 18F Tri-Flo suction catheter (Baxter Healthcare Corp., Valencia, CA) placed in the anterior nares. In children older than 4 yr of age, a blunt-tipped Yankauer catheter (Sherwood Medical) was used to externally occlude the nares for atraumatic suction. Afterwards, each catheter was rinsed with 1 ml of saline to rinse adherent secretions into the trap. The final volume combined from both nostrils ranged between 2.5 and 3 ml.

On a glass slide, a Hansel stain (Lide Laboratories, Inc., Florissant, MO) of nasal secretions obtained from the wash specimen was performed in order to assess eosinophil number (1). The remaining wash specimen was stored on ice for as long as 30 min, and then transferred using disposable pipettes to a plastic conical tube. After vortexing to achieve a homogeneous suspension of wash fluid and secretions, 0.5-ml aliquots were transferred by sterile pipette to 2-ml vials of viral transport media (Bartels Diagnostic, Inc., Issaquah, WA) and to a 1-ml sterile vial for viral culture and RT-PCR, respectively. The remaining wash specimen was then centrifuged at 1,000 × g for 5 min, and the supernatant was stored together with the RT-PCR specimen at 70° C for subsequent analyses.

Five milliliters of blood was obtained by venipuncture from each patient and processed as previously described (14). To determine total blood eosinophil number, 0.5 ml was sent for complete blood cell counts and differential cell analyses using an H*1 Blood Cell Analyzer (Technicon Instruments Corp., Tarrytown, NY).

Immunoassays

ECP in nasal wash fluid and serum and total serum IgE were measured using monoclonal antibody-based fluorometric assays (Pharmacia Cap System) provided by Pharmacia Diagnostics (Uppsala, Sweden) (14). Allergen-specific IgE antibody was measured by quantitative RAST (1). Each serum specimen was tested separately for dust mites (Dermatophagoides pteronyssinus and Dermatophagoides farinae), cat epithelium, cockroach (mixture of German, American, Oriental), short ragweed, and rye grass.

Viral Detection

RT-PCR for rhinovirus was performed on nasal wash samples using primers recognizing conserved picornavirus sequences, and the products were detected by microplate hybridization using an oligonucleotide probe specific for picornavirus (15). The specificity of this RT-PCR for picornavirus has been demonstrated against RSV, adenovirus, and coronavirus (16). Of nasal samples positive for picornavirus by RT-PCR, previous studies have confirmed that almost all positive isolates were rhinovirus (3) and that enterovirus-specific probes did not hybridize with the products (15). Similarly, RT-PCR for human coronavirus was accomplished using primers and oligonucleotide probes specific for human coronavirus OC43 and 229E (15). Duplicate monolayers of Ohio HeLa, human embryonic lung fibroblast (WI-38 strain), A549, and primary rhesus monkey kidney cell lines were inoculated with the 0.2 ml of nasal wash sample in viral transport media to isolate respiratory viruses using standard techniques. The two specimens that cultured enterovirus, a picornavirus, were positive by RT-PCR but not characterized in analyses as rhinovirus positive. RSV antigen in nasal wash samples was detected using a membrane enzyme immunoassay (Abbot Diagnostics Division, Chicago, IL).

Statistical Analysis

Data from wheezing and control patients were compared in two age groups: children younger than 2 yr of age, and children 2 to 16 yr of age. Chi-square analyses were used to compare the prevalence of positive viral tests, both as a single risk factor and in combination with IgE and eosinophil findings. When samples were small, Fisher's exact test was employed. Student's two-sample t tests were used to compare continuous variables between wheezing and control patients. Data not uniformly distributed were logarithmically transformed before t test analyses.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Viral Infection

Infants (younger than 2 yr of age). Respiratory viruses were detected in 82% (18 of 22) of wheezing infants younger than 24 mo of age. RSV was the most common viral pathogen, as recognized by a positive antigen test in 68% (15 of 22) of the nasal washes from infants in this wheezing group (Figure 1). RSV was isolated by culture from 11 of the 15 antigen-positive wheezing infants. All wheezing patients with a negative RSV antigen test also had a negative culture, and no control infant had a positive test for RSV either by culture or by antigen assay. Rhinovirus was detected by RT-PCR in 41% of both wheezing (9 of 22) and control (7 of 17) infants, and was cultured from the nasal washes of 23% of wheezing infants and 25% of control subjects (Figure 1). All positive rhinovirus cultures were accompanied by a positive RT-PCR test except one. Of the seven wheezing infants who were RSV antigen negative, three had a positive test for rhinovirus. The youngest of these who was positive for rhinovirus (both RT-PCR and culture) was 4 mo of age. RSV antigen and rhinovirus by RT-PCR were detected together in six wheezing infants, but in none of the control subjects. Other common viral pathogens were identified in two control subjects (adenovirus, human coronavirus OC43) and in one wheezing infant (enterovirus).

View larger version (28K): [in this window] [in a new window]   Figure 1.   Viral identification in nasal washes by age group. The height of each bar indicates the percentage of nasal washes positive for virus. Horizontally hatched bars indicate washes positive for RSV antigen. Closed and shaded bars represent washes positive for rhinovirus by RT-PCR and culture, respectively. Diagonally hatched bars indicate washes that tested positive for human coronavirus by RT-PCR or for other viruses in culture. ***p < 0.005 for wheezing versus control subjects in age group.

Children 2 to 16 yr of age. After the age of two, 83% of wheezing children (40 of 48) had evidence of viral infection. Rhinovirus was the predominant pathogen, as detected in nasal washes by RT-PCR in 34 of 48 (71%) wheezing children older than 2 yr of age compared with 15 of 42 (36%) of the nonwheezing control subjects (p < 0.005) (Figure 1). By comparison, rhinovirus was recoverable by culture in 18% of wheezing subjects and 24% of control subjects (p > 0.05). All nasal washes giving a positive rhinovirus culture except one were also positive by RT-PCR. RSV was detected in only 6% (3 of 48) of wheezing children and in three of the control subjects (p > 0.05). Of the three wheezing children positive for RSV, one (age 13) was older than 4 yr of age. Other viruses were also uncommon. Influenza A and enterovirus were isolated each from one wheezing child, and adenovirus was cultured from one control subject. Human coronavirus was detected by RT-PCR in three wheezing children (6%), two of whom had evidence of other viral infection by culture (influenza A and rhinovirus).

Relationship of Viral Infection to IgE and Eosinophil Findings

Infants (younger than 2 yr of age). We previously reported increased concentrations of ECP (> 200 ng/ml) in nasal washes from 41% of the wheezing infants and in one control subject from this study population (14). The prevalence of elevated nasal ECP levels among wheezing infants was higher in RSV positive (47%) than in RSV negative (29%) subjects, but not significantly so (p = 0.08). One of the two RSV negative wheezing infants with elevated nasal ECP was also rhinovirus positive by RT-PCR. None of the wheezing or control infants had IgE antibody to common aeroallergens by RAST, and only one infant (from the wheezing group) had nasal eosinophilia. Total IgE levels were low in this age group, although wheezing infants had higher levels than did control subjects (geometric mean = 7.0 and 3.7 IU/ml, respectively, p < 0.05).

Children 2 to 16 yr of age. The presence of IgE antibody to aeroallergens, nasal eosinophilia, and nasal ECP concentrations over 200 ng/ml were each strongly associated with wheezing (Table 1). Of these, the presence of 200 ng/ml or greater of ECP in nasal wash fluid gave the largest odds ratio (8.0) for wheezing. A positive RT-PCR for rhinovirus was also strongly associated with wheezing, but the odds ratio for wheezing was substantially increased among those children who had a positive RT-PCR test along with a positive RAST for IgE antibody (Table 1). The highest odds ratios for wheezing were observed in the children with a positive RT-PCR for rhinovirus together with nasal eosinophilia or elevated nasal ECP, a combination that was observed in only one control patient (Table 1). Forty-eight percent (23 of 48) of wheezing children had a positive test for rhinovirus and at least one positive nasal eosinophil marker (elevated nasal ECP or nasal eosinophilia), compared with 5% (2 of 42) of control subjects.

Among those children 2 yr of age and older who were rhinovirus positive by RT-PCR, wheezing subjects had higher levels of serum ECP (p < 0.001) and a higher percentage had nasal eosinophilia or elevated nasal ECP (p < 0.01) compared with control children (Table 2). Markers of atopy were more common in the rhinovirus negative wheezing patients (p < 0.05): 94% had at least one positive RAST (10 of 14) or total IgE greater than 200 IU/ml (10 of 14). In contrast, 59% of the rhinovirus positive wheezing subjects had a positive RAST (14 of 34) or elevated total IgE (16 of 34). Among the wheezing children with positive tests for rhinovirus and serum IgE antibody by RAST, significantly higher levels of total IgE and a larger percentage of nasal smears with eosinophilia were evident compared with those who were rhinovirus positive but RAST negative (p < 0.01) (Table 2). Systemic assessments of eosinophils in blood and ECP in serum, however, were very similar in the RAST positive and negative wheezing patients who had positive tests for rhinovirus.

Clinical Outcomes

Of the 22 wheezing infants younger than 2 yr of age, 45% were administered systemic corticosteroids (by mouth, intramuscularly, or intravenously) and 27% were admitted to the hospital. Steroid requirements did not associate with viral status (data not shown), but all hospitalized infants were RSV positive.

The majority (71%) of wheezing children older than 2 yr of age required systemic corticosteroids as part of their emergency visit, and 17% were hospitalized. Rhinovirus was commonly detected regardless of steroid (79% yes, 50% no) or admission (75% yes, 70% no) outcomes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This investigation indicates that most wheezing toddlers and older children needing emergency care have rhinovirus infection. Our observations are unique in that molecular techniques of viral detection were used to define the presence of rhinovirus in relationship to age, other common cold viruses, and to markers of atopy and airway inflammation. The pivotal role that rhinovirus appears to play in childhood wheezing illnesses has been underrecognized using cell culture detection as employed in our earlier emergency room investigation of a different group of wheezing children (1). Our present results also reaffirm the strong connection between RSV and wheezing before 2 yr of age and establish that coronavirus and other common respiratory viruses are much less likely to be present in older wheezing children in comparison with rhinovirus. This identification of rhinovirus as a dominant risk factor for wheezing is consistent with the temporal relationship seen between emergency visits for childhood wheezing at our institution and the peak seasons of rhinovirus activity (fall and late spring) in our community (17). The surge in wheezing visits to our emergency department in the autumn resulted in increased enrollment (37% of wheezing subjects older than 2 yr of age) to the study during September and October. The proportion of wheezing children, however, with a positive RT-PCR for rhinovirus during this period (67%) was similar to that found during the remainder of the year (74%). These prevalence results for rhinovirus are consistent with that observed in a prospective investigation of 9- to 11-yr-old children with asthma who experienced lower respiratory tract symptoms and decreased peak flows (3).

Our emergency department study conducted during the school year also detected rhinovirus in 40% of infants and in one-third of older children without cold symptoms. These observations suggest that there is a large pool of rhinovirus within the childhood community unaccompanied by overt clinical symptoms. To our knowledge, no prior investigation has examined nonwheezing children for rhinovirus by RT-PCR at the same time and in the same setting as wheezing subjects. In their community study of 9- to 11-yr-old asthmatics, Johnston and colleagues (3) found a lower prevalence of rhinovirus by RT-PCR (12%) in asymptomatic adolescents, but those control children were enrolled during a period when rhinovirus infections are likely at their nadir, i.e., the summer. Whether the positive tests for rhinovirus in the control children from our study represent resolving colds or subclinical infections is unknown because it is not clear how long a RT-PCR test remains positive for rhinovirus after naturally occurring infection. The pervasiveness of rhinovirus infection in children is further supported by its culture isolation from both wheezing and control subjects in our investigation. Positive culture for rhinovirus, however, did not distinguish wheezing school-aged children from control children despite the marked difference in rhinovirus detection by RT-PCR. This disparity may be a consequence of requiring infectious virus for culture recovery, whereas RT-PCR can recognize small amounts of viral RNA. Nasal wash samples from the subjects were obtained and stored on ice for as long as 30 min before processing. In contrast, nasal secretions were aspirated directly into viral transport medium at the time of enrollment in our previous emergency room study that demonstrated a positive association between rhinovirus culture isolation and wheezing (1).

The mechanism behind the close association between rhinovirus infection and wheezing attacks remains unclear. Rhinovirus has been detected in the lower airways after experimental infection (18), but it is not known to cause cytopathic changes in the respiratory mucosa (19). Instead, rhinovirus appears to influence airway responses that are unique to asthma. Experimental rhinovirus infection of adults with asymptomatic or mild asthma revealed increases in airway reactivity to histamine or methacholine but failed to show significant spontaneous small airway obstruction (12, 20). These investigations of asthmatic adults did not specifically examine subjects with evidence of ongoing atopic illness, but studies of experimental rhinovirus (serotype 16) infection in adults with allergic rhinitis have demonstrated enhanced early- and late-phase lower airway responses to allergen challenge (10, 11), and increased airway hyperreactivity compared with nonallergic control subjects (23). The importance of natural rhinovirus infection in concert with allergic inflammation is supported by our data revealing that the odds ratio for wheezing was significantly increased among children 2 to 16 yr of age who had the combination of a positive RT-PCR test for rhinovirus and the presence of serum IgE antibody to common aeroallergens by RAST. To a greater extent, these observations reprise the viral culture and RAST results seen in an investigation of a previous group of wheezing children from this institution (1). In reference to rhinovirus, IgE antibodies, and wheezing, it is intriguing to note that although rhinovirus was common (41%), it was not a risk factor for wheezing in infants, all of whom were RAST negative.

Despite the strong association with wheezing, rhinovirus in combination with a positive RAST characterized less than a third of the symptomatic children older than 2 yr of age. Many of the wheezing children could be identified by a positive test for rhinovirus or evidence of atopy alone, attesting to a heterogeneity of phenotypes for wheezing illnesses. In particular, markers of atopy were present in nearly all (13 of 14) of the wheezing children older than 2 yr of age who were rhinovirus negative. Eosinophilic airway inflammation is pathognomonic for asthma, and it is informative that the strongest risk profile for wheezing among children in our study was a positive RT-PCR test together with eosinophilia or elevated ECP in their nasal washes. A large proportion of these wheezing subjects with rhinovirus infection and elevated markers of eosinophil activity were RAST negative, and this could be taken to imply that rhinovirus infection may incite eosinophil/ECP responses in both atopic and nonatopic children with acute wheezing illnesses. Indeed, acute eosinophilic infiltrates of the bronchial mucosa have been demonstrated in normal control subjects as well as in atopic asthmatics after experimental rhinovirus infection (12). The concept of rhinovirus augmenting eosinophil responses is supported by the recent in vitro observations of Gern and colleagues (24) who found monocyte-driven T-cell activation after incubation with rhinovirus that was productive of factors that enhance eosinophil survival.

Elevated nasal ECP was also strongly associated with wheezing in children younger than 2 yr of age enrolled in this study (14). Most of the wheezing infants with elevated ECP in their nasal wash supernatants had positive tests for RSV. A role for RSV in the pathogenesis of an eosinophil response has been proposed based on the detection of IgE antibody specific for RSV in nasopharyngeal secretions from infants with bronchiolitis (13). However, IgE to RSV and markers of eosinophil inflammation (e.g., ECP) in nasal secretions have not been measured in the same patients. In animal studies, mice immunized with the capsular RSV-G attachment protein developed an eosinophil infiltrate in their lungs after subsequent intranasal inoculation of RSV (25, 26). In addition, more serious RSV disease associated with pulmonary eosinophilia has been observed in infants immunized with a formalin-inactivated RSV vaccine in the 1960s (27). Whether RSV can stimulate an eosinophil response in the respiratory tract during the course of natural infections is not yet certain, but it is clearly of interest in relation to studies suggesting that markers of eosinophil degranulation such as ECP might be predictive of early childhood asthma (28).

In summary, this investigation has demonstrated that the majority of emergency childhood wheezing illnesses can be defined by the presence of two respiratory viruses: RSV in infants before 2 yr of age and rhinovirus in older children. In addition, rhinovirus was detected in a large number of infants and children without apparent respiratory symptoms, implying that rhinovirus infection may be more pervasive during childhood than previously recognized. In the setting of emerging information that both RSV and rhinovirus are linked to the augmentation of proinflammatory mediators seen in wheezing illnesses, this study reveals that most infants with RSV-associated wheezing have evidence of eosinophil inflammation, and that the strongest odds of wheezing occur in older children who have rhinovirus together with evidence of atopy or eosinophil activity. The important clinical implications of this study are that efforts to prevent and treat wheezing attacks in children should focus on RSV and rhinovirus and their relationship to IgE and eosinophil-driven inflammatory responses.