Eosinophilia at the Time of Respiratory Syncytial Virus Bronchiolitis Predicts Childhood Reactive Airway Disease

(1)

Eosinophilia at the Time of Respiratory Syncytial Virus Bronchiolitis

Predicts Childhood Reactive Airway Disease

Daryl R. Ehlenfield, MD; Kathleen Cameron; and Robert C. Welliver, MD

ABSTRACT. Objective. Bronchiolitis in infancy is viewed as a risk factor for childhood asthma, but factors predicting which infants will have persistent wheezing have not been identified. In addition, the nature of the association between the 2 conditions is uncertain. We wished to determine whether eosinophil counts at the time of acute bronchiolitis predicted the presence of wheezing in later childhood.

Methods. We retrospectively identified infants hospi-talized with bronchiolitis, determined peripheral blood eosinophil counts at the time of bronchiolitis, and then contacted their families when they had reached 7 years of age.

Results. Eosinophil counts at the time of bronchioli-tis were greater in subjects who would have wheezing at 7 years of age (median: 98 cells/mm3_{) than in infants who}

would have no recurrent wheezing (median: 0 cells/mm3₎

or transient wheezing only up to 3 years of age (median: 0 cells/mm3_{). When the effects of family history of}

asthma, gender, and passive exposure to cigarette smoke were examined, only eosinophilia at the time of bronchi-olitis demonstrated a statistically significant relationship to the presence of wheezing at 7 years of age.

Conclusions. Eosinophilia at the time of bronchiolitis generally predicts the development of wheezing persist-ing into later childhood. Therefore, the association of bronchiolitis and childhood asthma seems more likely to be attributable to an immunologic anomaly that precedes the development of, or is induced by, bronchiolitis rather than to structural damage to the airway as a result of bronchiolitis.Pediatrics2000;105:79 – 83;respiratory syn-cytial virus, bronchiolitis, childhood asthma, eosinophils.

ABBREVIATIONS. RSV, respiratory syncytial virus; ECP, eosino-phil-cationic protein; LRI, lower respiratory illness; IL, interleukin; IFN-g, interferon-g; MIP-1a, macrophage inflammatory protein-1a.

R

espiratory syncytial virus (RSV) bronchiolitis in infancy is frequently complicated by the development of recurrent wheezing in later childhood.1_{No factors have been identified that can}

predict accurately which infants will have recurrent wheezing. Also, it has not been clearly determined whether bronchiolitis in infancy causes long-term airway changes, which result in childhood asthma,

or whether the occurrence of bronchiolitis indicates the presence of an underlying predisposition to wheezing both in infancy and in later childhood.

A previous study from this institution evaluated peripheral blood eosinophil counts in infants over 2 months of age with RSV infection.2_{That study}

dem-onstrated that infants of this age with upper respira-tory tract illness alone at the time of RSV infection, and most infants with RSV bronchiolitis, had sup-pression of eosinophil counts in comparison to healthy infants and those with noninfectious ill-nesses. However a subset of infants with bronchioli-tis, particularly males, had eosinophilia during acute bronchiolitis. Because elevated eosinophil counts are associated with childhood asthma, we undertook the present study to determine whether those infants who had detectable eosinophil counts during bron-chiolitis were more likely to develop wheezing that persisted into later childhood.

PATIENTS AND METHODS Study Subjects

The study was approved by the institutional review board of the Children’s Hospital of Buffalo. The population studied con-sisted of 43 infants hospitalized in the RSV seasons of 1991 and 1992 with acute bronchiolitis. Because infants,2 months of age generally have elevated eosinophil counts and maintain the ele-vation regardless of the form of RSV infection,2_{only infants 2 to 18} months of age were included. Infants with chronic underlying illnesses (including but not restricted to bronchopulmonary dys-plasia, cystic fibrosis, congenital heart disease, and immunodefi-ciency states) were excluded from analysis, as were infants receiv-ing corticosteroid therapy. All subjects were experiencreceiv-ing their first episode of wheezing, and all had RSV infection confirmed by the detection of RSV antigen in nasopharyngeal secretions using commercial direct immunofluorescence assays. Demographic fea-tures of study subjects are summarized in Table 1. Peripheral blood eosinophil counts were determined on the day of hospital-ization. Both the percentage of eosinophils and the total numbers of eosinophils (percentage of eosinophils multiplied by the total white blood cell count) were recorded. In most cases, only a single estimation of eosinophil counts was available. In 4 cases, a second count was obtained, and varied by,10% from the original count.

Follow-up

Infants with RSV bronchiolitis were initially identified by con-sulting records of the virology laboratory. Medical records per-taining to the initial hospitalization were then reviewed to confirm the diagnosis of bronchiolitis and to determine eosinophil counts at the time of bronchiolitis. A total of 104 patients were identified who met the age criteria, and in whom complete blood counts and differentials had been performed. Once the subjects had reached 7 years of age, we attempted to contact parents or guardians by telephone. Responses were obtained from 43 (41.3%) families, the other 61 families having moved away or being otherwise unavail-able. Parents then were asked whether the study subject had experienced any wheezing after the initial episode of bronchiolitis From the Department of Pediatrics, State University of New York at Buffalo,

School of Medicine and Biomedical Sciences, and the Division of Infectious Diseases, Children’s Hospital of Buffalo, Buffalo, New York.

Received for publication May 18, 1999; accepted Sep 29, 1999.

Reprint requests to (R.C.W.) Children’s Hospital, 219 Bryant St, Buffalo, NY 14222. E-mail: rwelliver@upa.chob.edu

(2)

and at what age these episodes of recurrent wheezing had taken place. Offices of private physicians or emergency department records were consulted to confirm parental reports. Passive expo-sure to cigarette smoke and family history of asthma were docu-mented in the medical record at the time of hospitalization for bronchiolitis and were determined again by telephone contact at the 7-year follow-up. All parents who reported smoking at the time of bronchiolitis had continued to smoke until the time of the follow-up telephone call, except in 1 case.

Definitions

Bronchiolitis was defined as the occurrence of wheezing of new onset at the time of RSV infection. All subjects were examined a minimum of 3 times during the hospitalization for bronchiolitis. Bronchiolitis also was considered to be present in the absence of wheezing if infants had hypoxia (oxygen saturation ,97% as determined by oximetry) and hyperinflation on chest radiographs. A positive family history of asthma was defined as the presence of an immediate family member who had received bronchodilator treatment for asthma. Exposure to cigarette smoke was defined as having taken place if someone smoked at least 3 cigarettes per day in the home of the study subject. Exposure to cigarette smoke in day care centers could not be evaluated. Recurrent wheezing was classified as either transient or persistent; in all cases, recurrent wheezing was accepted as having occurred only if the subject had received bronchodilators for the wheezing episode. Transient wheezing after bronchiolitis was defined as wheezing occurring as late as 3 years of age but not later. Persistent wheezing was defined as the occurrence of wheezing within the year preceding the telephone call.

Statistical Analysis

Because eosinophil counts were not randomly distributed, com-parisons of total eosinophil counts among groups were conducted using the nonparametric Mann-WhitneyUtest. Comparisons of groups with any or no eosinophilia (actually$1% vs,1% eosin-ophilia) and other comparisons of nominal variables were made using the Fisher’s exact test. Statistical comparisons were com-pleted using StatView 4.5 for Macintosh (Abacus Concepts, Berke-ley, CA).

RESULTS

Determinants of Eosinophil Counts During Bronchiolitis

Peripheral blood eosinophil counts at the time of bronchiolitis are analyzed by family history of asthma, gender, and passive exposure to cigarette smoke in Fig 1. Infants from asthmatic families had a median of 49 eosinophils/mm3 _{(range: 0 –352),}

whereas infants from nonasthmatic families had a median of 0 eosinophils/mm3 _{(range: 0 – 628;} _P 5

.52). Male infants had a median of 84 eosinophils/ mm3_{(range: 0 – 628), whereas females had a median}

eosinophil count of 0 cells/mm3_{(range: 0 –146;}_P ₅

.021). Infants exposed to cigarette smoke also had a median eosinophil count of 49 cells/mm3 _(range:

0 – 628), whereas infants not exposed had a median count of 0 cells/mm3_{(range: 0 – 410;}_P₅_.52).

Relationship of Eosinophil Counts and Recurrent Wheezing After Bronchiolitis

Figure 2 illustrates the eosinophil counts at the time of bronchiolitis in subjects who would later have either no wheezing episodes after bronchiolitis, transient wheezing through 3 years of age but not thereafter, or persistent wheezing through 7 years of age. Most infants who had no additional wheezing had no eosinophilia at the time of bronchiolitis. Only 5 of 23 (21.7%) of these subjects had eosinophilia, with a median of 0 cells/mm3_{. Of the 5 subjects who}

would have transient wheezing, 2 (40%) had eosin-ophilia with bronchiolitis (median: 0 cells/mm3_).

Fif-teen subjects had persistent wheezing at 7 years of age; 9 (60%) of these had eosinophilia at the time of bronchiolitis (median: 98 cells/mm3_{). Eosinophil}

counts at the time of bronchiolitis were greater in those subjects who would go on to have persistent wheezing than in those who would have no wheez-ing after bronchiolitis (P5.030). In addition, eosin-ophil counts were greater in subjects with persistent wheezing than in the group of subjects who were not wheezing at 7 years of age (groups with no wheezing and transient wheezing combined;P5 .046).

Factors Associated With Persistent Wheezing at 7 Years of Age

Figure 3 compares the effects on wheezing at 7 years of age of several factors recognized to be asso-ciated with childhood asthma. Neither passive smok-ing (P 5 .47) nor gender (P . .99) were associated with wheezing at 7 years of age. Of 12 infants from asthmatic families, 7 (58%) had wheezing at 7 years of age, whereas 8 (29%) of 31 children from nonas-thmatic families also had wheezing at 7 years of age, a difference that falls just short of statistical signifi-cance (P5.074). The only factor that was associated in a statistically significant fashion with wheezing at 7 years of age was the presence of eosinophilia at the time of bronchiolitis. That is 9 of 16 (56%) subjects who had eosinophilia with bronchiolitis had persis-tent wheezing, whereas only 6 of 27 (22%) subjects without detectable eosinophilia during bronchiolitis had persistent wheezing at 7 years of age (P5.045).

TABLE 1. Demographic Features of Study Subjects

Factor Number of Subjects in Group With Total

Subjects No Recurrent Wheezing Transient Wheezing Persistent Wheezing

Males 11 5 9 25

Females 12 0 6 18

Family history of asthma 3 2 7 12

Negative family history of asthma 20 3 8 31

White 17 5 11 33

Black 5 0 3 8

Hispanic 1 0 1 2

Passive smoking 5 3 8 16

(3)

Predictive Value of Eosinophilia for Persistent Wheezing

The positive predictive value of eosinophilia for wheezing at 7 years of age was 56%. The negative predictive value of the absence of eosinophilia for the absence of wheezing at 7 years of age was 78%. The sensitivity of eosinophilia during bronchiolitis for detection of subjects who would have persistent wheezing was 60%, whereas the specificity was 75%.

DISCUSSION

The results of this study demonstrate that infants who have peripheral blood eosinophilia at the time of RSV bronchiolitis are more likely to have wheez-ing that persists until 7 years of age than are infants who are capable of suppressing eosinophilia during bronchiolitis. The association of eosinophilia with persistent wheezing could not be explained based on gender, family history of asthma, or exposure to

Fig 1. Factors associated with peripheral blood eosinophilia during bronchiolitis. Bars indicate median eosinophil counts in infants with bronchiolitis. Male infants with bronchiolitis had higher eosinophil counts (median: 84/mm3_{) than female infants (median: 0 cells/mm}3_). Eosinophil counts did not differ significantly in infants with a family history of asthma, compared with infants from nonasthmatic families. Eosinophil counts in infants exposed to cigarette smoke in the home were not different from those of infants from nonsmoking families.

(4)

cigarette smoke in the home, each of which has been associated with the development of childhood asthma. Eosinophilia with bronchiolitis was not ex-plained by a family history of asthma. The risk for persistent wheezing was determined using any eo-sinophilia (in fact,$1% eosinophilia) as the cutoff for analysis. This corresponded to an eosinophil count of

$84 cells per mm3_{, the lowest calculated count}

ob-served.

Three previous studies have evaluated the rela-tionship of the quantity of eosinophil-cationic pro-tein (ECP), a product of eosinophil degranulation, in serum at the time of bronchiolitis to the development of wheezing immediately after bronchiolitis with conflicting results. Reijonen and colleagues3_followed

infants for 16 weeks after an acute episode of bron-chiolitis and found that recurrent wheezing during this brief interval was more common in infants who had higher serum ECP concentrations at the time of bronchiolitis. Koller et al4_{followed infants with}

bron-chiolitis for 12 months after acute bronbron-chiolitis and similarly found an association of greater ECP con-centrations at the time of bronchiolitis with more frequent recurrences of wheezing during this limited interval. Neither of these studies followed subjects into later childhood. This is important in that many infants who have recurrent wheezing shortly follow-ing acute bronchiolitis will no longer have wheezfollow-ing after 2 to 3 years of age.5 _{Thus, the relationship of}

ECP concentrations during bronchiolitis to child-hood asthma could not be assessed in these studies of ECP. In a third study, Oymar and Bjerknes6_found

no association of serum ECP at the time of bronchi-olitis to the presence of recurrent wheezing in pa-tients followed for an average of 3 years. The longer period of follow-up in the present study indicates that the association of eosinophil activation during bronchiolitis and recurrent wheezing continues at least through 7 years of age.

Martinez and colleagues7_{studied peripheral blood}

eosinophil counts in infants at the time of their first episode of lower respiratory illness (LRI). These in-fants were then followed through 6 years of age. As in the present study, higher eosinophil counts at the

time of initial LRI were associated with wheezing persisting to 6 years of age, but not with wheezing that ceased at 3 years of age. Therefore, the findings of the present study confirm those of Martinez and colleagues and extend them in several ways. First, only approximately two thirds of the subjects of the study by Martinez et al were diagnosed with bron-chiolitis at the time of first LRI. Other subjects had an initial diagnosis of croup (10%–12%), pneumonia (4%– 8%), bronchitis (6%–10%), or other respiratory conditions (3.5%–7%). Second, the study by Martinez et al did not determine whether eosinophilia was related to a family history of atopy or to gender. Third, their subjects were seen as outpatients and had a variety of viral agents identified as the cause of their initial LRI episodes. The subjects in the present study all had bronchiolitis attributable to RSV infec-tion as the initial respiratory illness, and all were hospitalized. Therefore, the association of eosino-philia and persistent wheezing holds even in those infants with more severe forms of bronchiolitis, in which persistent airway damage as a result of the infection itself was perhaps more likely to occur.

The mechanism by which eosinophilia is main-tained in those infants with bronchiolitis who will have persistent wheezing has not been investigated. Secretion of increased amounts of interleukin- (IL-) 4 and IL-5, which are chemotactic for eosinophils,8,9

represents 1 possible mechanism. However quanti-ties of these 2 cytokines do not seem to be increased in bronchiolitis.10 –13 _{In fact, the results of a study}

from this institution suggested that interferon-g (IFN-g), which is believed to oppose the activity of IL-4 and IL-5, was the predominant cytokine re-leased into the airway during RSV-induced wheez-ing illnesses.13 _{In a more recent study (manuscript}

submitted), we found that macrophage inflamma-tory protein-1a (MIP-1a), a chemokine that is also chemotactic for eosinophils, is present in higher con-centrations in the respiratory secretions of infants with RSV bronchiolitis than in secretions from healthy infants or those with upper respiratory ill-ness alone attributable to RSV infection. The release of MIP-1a is enhanced by IFN-g. Therefore, certain

(5)

infants may respond to RSV infection by releasing particularly greater quantities of IFN-g, and subse-quently, MIP-1a. These individuals might then have higher eosinophil counts during bronchiolitis. It is also possible that eosinophils might be absent from the blood of infants with bronchiolitis because they have migrated into the lung or airway lumen. This is unlikely, because eosinophils have not been ob-served in lung tissues of infants with fatal RSV in-fection15 _{or in samples of nasopharyngeal secretions}

examined by ourselves (unpublished observations) or others.16_{A third possible explanation for our}

find-ings is that infants with bronchiolitis have a reduced capacity for suppressing eosinophil counts during viral infections. There is no evidence available cur-rently to confirm or dispute this possibility. An al-ternative explanation might be that infants with eo-sinophilia at the time of bronchiolitis have a reduced capacity for suppressing eosinophil counts during viral infections.

The association of eosinophilia with recurrent wheezing after bronchiolitis provides a general mea-sure for predicting which infants with comparatively severe forms of bronchiolitis (those hospitalized with the initial bronchiolitis episode) will have persistent wheezing in later childhood. The study by Martinez and colleagues7 _{involved mostly outpatients,}

sug-gesting that the association is true for infants with milder forms of bronchiolitis also. More importantly the results seem to indicate that an immunologic mechanism, which causes the association of eosino-philia and wheezing, is already present at the time of infantile bronchiolitis, as it is at the time of childhood asthma. Although additional proof is required, these findings suggest that the association of bronchiolitis in infancy with childhood asthma is not attributable to damage to the airway as a result of RSV infection, but rather to a common immunologic anomaly that is either induced by RSV infection or which was present before RSV infection occurred. These find-ings, therefore, add to our knowledge of how asthma develops in childhood.

REFERENCES

1. McConnochie KM, Roghman KJ. Bronchiolitis as a possible cause of wheezing in childhood: new evidence.Pediatrics. 1984;74:1–10 2. Garofalo R, Dorris A, Ahlstedt S, Welliver RC. Peripheral blood

eosin-ophil counts and eosineosin-ophil cationic protein content of respiratory secretions in bronchiolitis: relationship to severity of disease.Pediatr Allergy Immunol. 1994;5:111–117

3. Reijonen TM, Korppi M, Kuikka L, et al. Serum eosinophil cationic protein as a predictor of wheezing after bronchiolitis.Pediatr Pulmonol. 1997;23:397– 403

4. Koller D, Wojnarowski C, Herkner KR, et al. High levels of eosinophil cationic protein in wheezing infants predict the development of asthma. J Allergy Clin Immunol. 1997;99:752–756

5. Henderson FW, Clyde WA, Collier AM, et al. The etiologic and epide-miologic spectrum of bronchiolitis in pediatric practice.J Pediatr. 1979; 95:183–190

6. Oymar K, Bjerknes R. Is serum eosinophil cationic protein in bronchi-olitis a predictor of asthma?Pediatr Allergy Immunol. 1998;9:204 –207 7. Martinez FD, Stern DA, Wright AL, Taussig LM, Halonen M.

Differen-tial immune responses to acute lower respiratory illness in early life and subsequent development of persistent wheezing and asthma.J Allergy Clin Immunol. 1998;:102:915–920

8. Moser R, Fehr J, Bruijnzeel PLB. IL-4 controls the selective endothelium-driven transmigration of eosinophils from allergic individuals.J Immu-nol. 1992;149:1432–1438

9. Lopez AF, Sanderson CJ, Gamble JR, Campbell HD, Young IG, Vadas MA. Recombinant human interleukin 5 is a selective activator of human eosinophil function.J Exp Med. 1988;167:219 –224

10. Roman M, Calhoun WJ, Hinton KL, et al. Respiratory syncytial virus infection in infants is associated with predominant Th-2-like response. Am J Respir Crit Care Med. 1997;156:190 –195

11. Renzi PM, Turgeon JP, Yang JP, et al. Cellular immunity is activated and a Th-2 response is associated with early wheezing in infants after bronchiolitis.J Pediatr. 1997;130:584 –593

12. Oymar K, Elsayed S, Bjerknes R. Serum eosinophil cationic protein and interleukin-5 in children with bronchial asthma and acute bronchiolitis. Pediatr Allergy Immunol. 1996;7:180 –186

13. van Schaik SM, Tristram DA, Nagpal IS, Hintz KM, Welliver RC II, Welliver RC. Increased production of IFN-gand cysteinyl leukotrienes in virus-induced wheezing.J Allergy Clin Immunol. 1999;103:630 – 636 14. Dery RE, Bissonnette EY. IFN-gpotentiates the release of TNF-aand

MIP-1aby alveolar macrophages during allergic reactions.Am J Respir Cell Mol Biol. 1999;20:407– 412

15. Neilson KA, Yunis EJ. Demonstration of respiratory syncytial virus in an autopsy series.Pediatr Pathol. 1990;10:491–502

(6)

DOI: 10.1542/peds.105.1.79

2000;105;79

Pediatrics

Daryl R. Ehlenfield, Kathleen Cameron and Robert C. Welliver

Childhood Reactive Airway Disease