|Home | About | Journals | Submit | Contact Us | Français|
The relevance of allergic sensitization, judged by titers of serum IgE antibodies, to the risk of an asthma exacerbation caused by rhinovirus is unclear.
To examine the prevalence of rhinovirus infections in relation to the atopic status of children treated for wheezing in Costa Rica, a country with an increased asthma burden.
The children enrolled (n=287) were 7 through 12 years old. They included 96 with acute wheezing, 65 with stable asthma, and 126 non-asthmatic controls. PCR methods, including gene sequencing to identify rhinovirus strains, were used to identify viral pathogens in nasal washes. Results were examined in relation to wheezing, total IgE, allergen-specific IgE antibody, and levels of expired nitric oxide (FENO).
Sixty-four percent of wheezing children compared to 13% of children with stable asthma and 17% of the non-asthmatic controls tested positive for rhinovirus (p<0.001 for both comparisons). Among wheezing subjects, 75% of the rhinoviruses detected were Group C strains. High titers of IgE antibodies to dust mite allergen (especially Dermatophagoides sp) were common and correlated significantly with levels of total IgE and FENO. The greatest risk for wheezing was observed among children with titers of IgE antibodies to dust mite ≥17.5 IU/ml who tested positive for rhinovirus (odds ratio for wheezing: 31.5; 95% CI 8.3–108, p<0.001).
High titers of IgE antibody to dust mite allergen were common and significantly increased the risk for acute wheezing provoked by rhinovirus among asthmatic children.
In countries with temperate climates (e.g., North America, Europe, Australia), viral respiratory tract infections are associated with 80 to 90% of wheezing attacks in the pediatric population, especially during early childhood (1–4). After 3 years of age, rhinovirus (RV) accounts for 75 to 80% of the virus-induced attacks leading to hospitalizations and emergency room (ER) visits, and the majority of these children are atopic (1,5). Evidence to date also indicates that an RV infection together with sensitization and exposure to inhaled allergens increases the risk for acute symptoms, suggesting that these risk factors may act synergistically to provoke asthma exacerbations (1,5,6). Whether these observations can be generalized to children living in countries with tropical climates is not clear. In studies from Brazil and Trinidad the prevalence of infections with RV among children treated for asthma was less than half of what has been reported in countries with temperate climates (7,8). Recognizing that additional studies are needed from tropical environments to gain a better understanding for role of viral infections in the etiology of asthma exacerbations worldwide, the purpose of this investigation was to examine the prevalence of viral respiratory tract infections among children treated for acute wheezing in Costa Rica and to evaluate the results in relation to their atopic status.
Costa Rica is a small country (19,730 square miles) in Central America with a diverse ecosystem. Similar to other countries near the equator, Costa Rica has two seasons: a “dry” season (generally December through April) and a “rainy” season (May through November). Exposure to helminthes, which, like environmental allergens, can stimulate the production of IgE, is common in economically-deprived regions of Hispanic America. However, anti-helminth medications are given annually to children in Costa Rica starting at one year of age. Thus, active parasite and enteric infections during childhood are not frequent in that country (e.g., ≤ 2% for infections with A. lumbricoides) (9,10).
According to school surveys and studies using the ISAAC questionnaire, the prevalence of asthma among school-aged children in Costa Rica (~23–27%) is higher than in North America (11,12), and sensitization to allergens, especially to allergens produced by the house dust mite, is common (13). At present, however, there is little information about the relevance of viral infections to acute attacks of asthma in this country. Additionally, recent evidence from countries with temperate climates suggest that Group C strains of RV may be strongly related to asthma exacerbation compared to other strains (14–16). This information is also lacking in studies of asthma in tropical countries. Thus, our objective was to investigate the relationship between respiratory viruses, including different strains of RV, and episodes of asthma requiring acute treatment. Aditionally, we took advantage of the dominance of dust mite sensitization in Costa Rica to investigate the relationship of titers of IgE antibodies to exacerbations of asthma with or without evidence of a recent viral infection.
This was a cross-sectional, case-control investigation of 287 children, ages 7–12 years old, enrolled in the emergency room (ER) of the Hospital Nacional de Niños, the main tertiary care hospital for children in San José, Costa Rica where the majority of children are seen for trauma, wheezing illnesses, vomiting, diarrhea, abdominal pain, and fever. Subjects included 96 children referred to the acute care nebulization room for wheezing by a triage physician. They required at least a nebulized bronchodilator (albuterol) for treatment. Children seen in the ER for non-wheezing disorders (n = 191) were enrolled as controls. They included 65 children who had been hospitalized or treated in the ER, or who had been using medications prescribed by a physician for asthma during the last 12 months. Data from the latter group, defined as having “stable asthma” at the time of enrollment, were compared to the data obtained from the actively wheezing children and other non-wheezing controls in a post-hoc analysis. Children with chronic lung disease, congenital heart disease, or children with immunodeficiency or oncology disorders were not enrolled.
The subjects included 137 children (44 with wheezing) enrolled in February 2009 during the “dry” season when children in Costa Rica begin the school year, and 150 children (51 with wheezing) were enrolled in October 2009 during the “rainy” season, one month before the end of the school year. Demographic information and subject characteristics were obtained from questionnaires administered to parents. The questionnaires focused on each child’s history for asthma treatments, family history for allergic disorders, and tobacco smoke exposure (ETS) at home. Informed consent was obtained from parents and informed assent from children who participated. The study was approved by the Ethics Committee at the Hospital Nacional de Niños and by the IRB at the University of Virginia.
Nasal washes were obtained for viral analyses as described online. Initially, they were evaluated for RV by reverse transcriptase PCR as described previously (17,18). Other respiratory viral pathogens were evaluated using real-time PCR assays obtained from the Centers for Disease Control and Prevention according to published procedures (19,20). These assays included tests for RV as well as tests for influenza A (including H1N1) and B, respiratory syncytial virus, human metapneumovirus, parainfluenza viruses 1, 2 and 3, coronaviruses (229E, OC43, NL63, and HKU1 species), and adenovirus. A high degree of concordance was observed between reverse transcriptase and real-time PCR methods for detecting RV (% absolute agreement = 93.3%; 95% CI = 89.8–95.9%). Additionally, strains of RV and enterovirus were identified by PCR and sequencing of a region comprising the VP4 and partial VP2 capsid protein genes (21,22).
Blood (5 ml) was obtained by venipuncture and serum from each sample was analyzed for total IgE using the Phadia ImmunoCAP assay (Uppsala, Sweden). Each sample was also analyzed for allergen-specific IgE antibody to dust mite (D. pteronyssinus, D. farinae, and Blomia tropicalis), Alternaria, Aspergillus, cockroach (Periplaneta americana and Blattella germanica), Bahia grass, cat and dog allergens, and Ascaris lumbricoides. Sera with ≥0.35 IU/ml of IgE antibody to any of the allergens tested were considered positive for allergen sensitization. FENO was measured using a portable FENO analyzer (NIOX MINO, Aerocrine, Inc., New Providence, NJ).
Questionnaire data and frequencies for positive tests for viral pathogens and allergen sensitization were analyzed via robust exact binomial and exact multinomial contingency table methods. Binomial contingency table hypotheses were evaluated via the exact binomial test, while multinomial contingency table hypotheses were evaluated via the Pearson exact goodness of fit test. For both the binomial and multinomial contingency table analyses, the two-sided null hypothesis rejection rule was set at p≤0.05. Multivariate logistic regression was utilized to determine whether a child’s wheezing status was associated with RV infection and atopic status. Tests of association were based on the Type III Wald chi-square statistic, and a p≤0.05 criterion was utilized to identify significant associations. Total serum IgE levels, titers of allergen-specific IgE antibody, and FENO levels were analyzed on a logarithmic scale by 2-way ANOVA. The two sources of variation considered in the ANOVA were the “study-group”, and the “season” of data collection. The rejection rule for hypothesis testing was based on a p≤0.05 criterion and 95% CI construction for the ratio of the geometric means (GMs) was based on the Student t-distribution. The statistical software package “SAS version 9.2.2” (SAS Institute Inc., Cary, NC) was used to conduct the statistical analyses.
Among the children enrolled who presented to the ED with a diagnosis that did not involve breathlessness, 34% (65/191) had stable asthma, as judged by parental report of treatment regimens. The percentages of the 96 wheezing children and those with stable asthma who had required hospitalization or treatment in the ER, or who used medications (bronchodilator and/or controller) for asthma during the last 12 months were similar (Table I). A minority of children in this study were exposed to ETS at home (23%), more often from the father. Children with stable asthma had less exposure to ETS at home and they used inhaled and nasal steroids daily more often than children enrolled for wheezing (Table I). More detailed comparisons for children enrolled in February and October are shown online (Table E1).
The percentage of children testing positive for any virus was significantly greater among wheezing children than among those with stable asthma or non-asthmatic controls (Table II). The percentages were similar among wheezing children during both enrollment periods (84% in February and 83% in October). The percentage of those testing positive for RV was also significantly greater among the wheezing children during both enrollment periods, but greater among wheezing children enrolled in February (80% (35/44) than October (48% (25/51)(p<0.029) (Fig. 1). Based on real-time PCR results, RV accounted for 74% of the viral pathogens detected (Table II). RSV and enterovirus (predominantly strain 68) were the only other pathogens associated with wheezing (10.4% and 6.3%, respectively), and these two viral pathogens were only detected in washes from children who were enrolled in October (Fig 1 and Table II).
Gene sequencing revealed that 75% of the nasal washes testing positive for RV from wheezing children were positive for Group C and 25% were positive for Group A strains (Table II). The percentage of RVs identified as Group C was similar among wheezing children enrolled in February and October (71% and 82%, respectively, p= 0.35). Four of the 8 stable asthmatics with positive tests for RV also tested positive for Group C. By comparison, 6 and 8 of the 17 RVs detected in washes from non-asthmatic controls were Group C and A strains, respectively (p= 0.79). Only 6 children tested positive for Group B strains, one of whom was enrolled for wheezing. Additional information about the RV strains detected is shown online (Table E2).
The geometric means (GM) for total IgE in sera from wheezing children (482 IU/ml; 95% CI 363–640) and stable asthmatics (332 IU/ml; 95% CI 235–469) were not significantly different (p= 0.10), but both values were higher than total IgE levels from the non-asthmatic controls (86 IU/ml; 95% CI 67–111) (p<0.001). Ninety-six percent of the wheezing children, 84% with stable asthma, and 54% of the non-asthmatic controls had at least one positive test for IgE antibodies to any allergen tested (p<0.001 for wheezing or stable asthmatic subjects compared to controls; p<0.05 for wheezing compared to stable asthmatics) (Table III). The most frequent IgE antibody responses were to dust mite allergens (i.e., D. pteronyssinus (Dp), D. farinae (Df), or Blomia tropicalis (Bt)), and a strong correlation was observed between positive tests for IgE antibodies to Dp and Df (rs = 0.95; 95% CI 0.94–0.96, p<0.001). Among the wheezing and stable asthma groups, both the prevalence of sensitization and the GM of IgE antibody titers to allergens from each of the mite species were significantly greater compared to results from the non-asthmatic controls (p<0.001)(Table III). In addition, when adjusted for sensitization to the other allergens, sensitization to dust mite was the only allergen that remained significantly associated with asthma (adjusted OR; 4.9, 95% CI 2.1–11.6 for IgE to Bt and 2.7, 95% CI 1.2–6.4 for IgE to Dp).
Sensitization to the cockroach species, Blattella germanica (Bg) and Periplaneta americana (Pa), and to the helminth, Ascaris lumbricoides, was also greater among the wheezing and stable asthmatic children compared to non-asthmatic controls (Table III). However, the prevalence of sensitization and titers of IgE antibodies (GM) to these allergens were much lower than values for any of the dust mite species. Additionally, a strong correlation was observed between total serum IgE levels and the titers of IgE antibodies to Df, Dp, or Bt among asthmatic subjects (wheezing and stable asthmatic children combined) (r=0.72, 0.75, and 0.69, respectively; each p<0.001). Data on other allergens and for IgE values in February and October are shown online (Table E3).
After adjusting for other viral pathogens and for specific IgE antibodies, the main pathogenic factor which remained significantly associated with wheezing among the children enrolled was a positive test for RV (adjusted odds ratio; 14.3; 95% CI 5.6–36.4, p<0.001). The odds for wheezing, however, were strongly influenced by the titer of IgE antibodies to dust mite allergen. Among 45 children with titers of IgE antibodies to Dp ≥ 17.5 IU/ml and a positive PCR test for RV, 42 children (93%) required treatment for acute wheezing (odds ratio: 31.5; 95% CI 8.3–108) compared to 16 of 23 children (70%) with titers of IgE antibodies between 0.35 and 17.4 IU/ml (odds ratio: 12.3; 95% CI 3.8–39) (Table IV). Moreover, the probability of being enrolled for acute wheezing, based on logistic regression analyses, was significantly associated with increasing titers of IgE antibodies to Dp (Fig. 2; panel A) and this risk was substantially higher among those who tested positive for RV (Fig. 2; panel B). The relationship between IgE antibody titers and Group C RV strains was the same as that for Group A RV (Table IV). In addition, despite smaller numbers, the other two viruses associated with wheezing (RSV and enterovirus) show the same trend (Table E4).
Levels of FENO (GM) were significantly higher among wheezing children (35 ppb; 95% CI 29–41) and stable asthmatics (25 ppb; 95% CI 21–31) compared to levels among non-asthmatic controls (14 ppb; 95% CI 12–16) (p<0.001 for both comparisons, and p<0.05 for wheezing compared to stable asthmatics). FENO levels were also higher among those who tested positive compared to those who tested negative for RV (40 compared to 28 ppb, p= 0.038), but were similar among those who tested positive for Group C and Group A strains (42 vs. 36 ppb, respectively, p= 0.52). A positive correlation was observed between levels of FENO and titers of IgE antibodies to mite in serum from children within each enrollment group (Fig 3). By contrast, FENO was not associated with sensitization to the cockroach species or Ascaris.
Infections with RV, most often Group C strains, were strongly associated with active wheezing among children (ages 7–12 years) living in Costa Rica. Most of the wheezing children (96% in this study) were atopic, and a large majority of them (88%) were sensitized to dust mite allergen. Moreover, the titers of IgE antibody to dust mite among children enrolled for wheezing, as well as those with stable asthma, were significantly greater than to any other allergen tested. Taken together, this is the first study to demonstrate that children with high titers of IgE antibodies, which in this study were predominantly to dust mite allergen, had the greatest risk for an attack of wheezing provoked by RV. Additionally, the results demonstrate that the pathogenic effects of individual strains of RV and other respiratory viruses may be difficult to determine without understanding the atopic characteristics of the host.
The rationale for enrolling subjects in the ER was that children could be evaluated without knowledge about their atopic status or whether they were recently infected with a viral pathogen. In this population, subjects enrolled for wheezing and those with stable asthma had similar allergen-specific IgE responses and their levels of total IgE and FENO were significantly increased compared to the non-asthmatic controls. After adjusting for atopic characteristics, exposure to ETS at home, and daily use of inhaled and nasal steroids, the main pathogenic factor which differentiated the children who required treatment for wheezing from those with stable asthma was a positive test for RV.
In previous studies from Brazil and Trinidad, RV infections were not associated with attacks of asthma as strongly as studies from countries with temperate climates have reported (7,8). Differences in lifestyle (e.g., quality of drinking water, rates of helminth infection, the amount of time spent indoors at home or in school) may explain differences in the rates of infection; however, both Costa Rica and the region of Brazil around Sao Paulo have developing Westernized lifestyles and a strong association between atopy and asthma was observed in both countries. Strains of RV within Groups A, B, and C were also evaluated for all positive tests for RV in our study. The majority of wheezing children were infected with Group C strains (i.e., 75% positive for Group C and 25% positive for Group A strains). Other studies of children treated for exacerbations of asthma have also reported a higher frequency of infection with Group C strains (14,15). Thus far, no studies have shown a significant association between asthma and Group B strains.
Information about the relationship between rates of infection with strains A, B and C in the general population is unclear. Data from the non-asthmatic controls in the present investigation, as well as from younger children evaluated for respiratory infections in other studies, suggest that the frequency of infections with Group A and C strains may be similar (16,21). This suggests that the airway in patients with asthma may be more susceptible to infections with Group C strains. At present, the receptor used by Group C viruses to infect epithelial cells is not known. The cell surface receptor used by Group A strains is the adhesion molecule, ICAM-1 (23,24). Previous studies, including our own, indicate that the expression of ICAM-1 (both soluble and membrane ICAM-1) is reduced in the asthmatic airway (25,26), which might give Group C viruses a selective advantage for infecting the asthmatic respiratory tract epithelium. In our patient population, FENO and specific IgE levels were similar among children infected with Group C and Group A strains. Additionally, the risk for wheezing with RV among children with high titers of IgE antibodies to dust mite allergen was not significantly increased among those infected with Group C strains. Recently, the significance of allergy in the response to RV was also suggested by the significant decrease in asthma exacerbations among children treated with omalizumab (27). The beneficial effects of inhaled corticosteroids, which also reduce Th2 inflammation and prevent asthma exacerbations during childhood, also support this observation (28,29).
High rates of sensitization to dust mite among children with asthma in Costa Rica have been described, and persistent humidity there favors mite growth throughout the year (13,30). Although measurements of allergen exposure in homes of the children enrolled were not done, previous investigations from Costa Rica suggest that exposure levels, based on measurements of guanine to estimate mite allergen concentrations in dust samples, are elevated throughout the year, but may be higher during the rainy season (30). Persistent humidity supporting mite growth is also present in New Zealand and the UK where high titers of IgE antibodies to dust mite and the risk of asthma have also been observed (31–33). In the present study, high titers of IgE ab to dust mite were strongly associated with asthma and significantly increased the risk for acute wheezing more than any other allergen tested. Most striking was the significant increase in the probability of wheezing with a recent RV infection among children with titers of IgE antibodies to Dp ≥17.5 IU/ml. Other environmental antigens capable of stimulating IgE antibodies may exist which were not evaluated. However, as judged by the prevalence and titers of IgE ab, none of the other allergens included in this study compared to the results observed for dust mite.
The observation that the generation of FENO in the asthmatic airway correlates with sensitization and exposure to inhaled allergens has been described previously, and levels of FeNO among the asthmatic children (both those with current wheeze and stable asthma) in the present study were significantly higher than among the non-asthmatic controls (32–34). These results are consistent with results from our experimental challenges with RV-16 demonstrating that the levels of FENO were significantly increased before the virus inoculation among asthmatics with high levels of total IgE (25). Together, the results suggest that pre-existing allergic inflammation is the main risk factor for an asthmatic response to RV. In the present study, the positive correlation between titers of IgE to Dp, Df, or Bt and FENO levels provides further evidence for the relevance of dust mite allergen to the pathogenesis of airway inflammation.
In conclusion, this investigation highlights the significant risk for asthma attacks caused by RV among children with high titers of IgE antibodies, which was most often associated with sensitization to dust mite allergen in Costa Rica. Thus far, anti-viral treatments, including vaccines, to treat infections with RV have been difficult to develop, because of antigenic diversity among RV strains. The predisposition of asthmatics to become infected with Group C strains might offer a more targeted approach for developing anti-viral treatments. However, this approach would not be effective for the 40% of wheezing children who did not test positive for RV in this study, and may not be effective in preventing attacks of asthma caused by Group A strains. Alternatively, treatments focused on decreasing allergic airway inflammation could provide a broader approach to prevent and/or decrease the severity of asthma exacerbations caused by RV. Thus, results from this study provide strong evidence that characterizing the atopic status of asthmatic children in greater detail, including their titers of allergen-specific IgE antibodies, is needed both to enhance investigations focused on mechanisms of RV-induced attacks of asthma and to identify children who are most likely to benefit from allergen-specific therapies.
High titers of allergen-specific IgE, in this study to dust mite, will identify children who may benefit most from therapies designed to prevent or treat rhinovirus-induced attacks of asthma.
Funding Sources: The Cove Point Foundation; NIH grants R01 AI020565 and U19 AI070364; University of Virginia Children’s Hospital Research Fund
We want to thank Lyn Melton for her time and help in preparing this manuscript.
Conflicts of Interest: None