INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Whether or not bronchial hyperresponsiveness (BHR) in children without a diagnosis of asthma is a risk factor for the development of asthma or asthma-like symptoms later in life is still a matter of debate. Although cross-sectional studies have demonstrated that BHR is associated with the presence of asthmatic symptoms in children (1), recent longitudinal studies (4) offered controversial answers to the question whether BHR can predict asthma. All these studies, however, were based on BHR tests that were performed in school-age children whose history of respiratory illness until the time of testing was either limited to the previous year (5, 7) or was obtained by retrospective questionnaires for the preschool years (4, 6, 8). It is thus not possible to exclude the possibility that respiratory symptoms compatible with a diagnosis of asthma could have been present in some of these children before they were tested for BHR.

Future strategies for the prevention of asthma will need to rely on the availability of markers that may allow for the identification of subjects at risk for the development of the disease. Understanding the potential role of BHR as a marker of a "pre-asthmatic state" is important not only to shed light on the pathophysiology of asthma but could also have important clinical applications.

The purposes of this study were (1) to assess BHR at age 6 yr in a group of children who did not have a diagnosis of asthma and who did not have a history of respiratory symptoms compatible with such a diagnosis and (2) to study the relation between BHR and the incidence of new cases of asthma or asthma-like symptoms during the subsequent years. Respiratory symptoms and asthma diagnosis were evaluated prospectively from birth and were thus not subjected to preferential recall bias at the time of BHR testing.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design

This analysis was part of the Tucson Children's Respiratory Study, a longitudinal study of respiratory illness involving children enrolled at birth between May 1980 and October 1984 (9, 10). All subjects were patients of the pediatricians of a large local health maintenance organization. At the time of enrollment, parents completed a questionnaire about their own respiratory symptoms, education, and smoking habits. Parents were also asked to take their children to the pediatrician whenever they had signs and symptoms of lower respiratory tract illnesses (LRIs) such as deep or "wet" chest cough, wheezing, hoarseness, stridor, or shortness of breath. Pediatricians assessed such signs and symptoms directly.

Parents were asked to complete a questionnaire about their children's respiratory symptoms and again about their own smoking habits at the following children's mean ± SD ages: 1.6 ± 0.3 yr (age 2 survey), 2.9 ± 0.5 yr (age 3 survey), and 6.0 ± 0.6 yr (age 6 survey). During the latter survey, BHR to isocapnic hyperventilation with cold dry air (CA) was assessed in the children, and skin tests for prevalent aeroallergens in the Tucson area were also performed. For the purpose of this study, children who had ever been diagnosed as having asthma by a physician by the time of the age 6 survey were excluded from the analysis. Children who had more than three episodes of wheezing during the 12 mo prior to the age 2, age 3, and age 6 surveys and those who had evidence of severe asthma (that is, who had an emergency room visit for wheezing during the 12 mo prior to the CA challenge or who used asthma medications for more than 21 d during the 6 mo prior to the CA challenge) were also excluded regardless of a diagnosis of asthma. After these exclusions, a total of 368 subjects were available for analysis.

At a mean ± SD age of 8.5 ± 0.7 yr (age 8 survey) and 10.8 ± 0.6 yr (age 11 survey), new information about the subjects' respiratory symptoms was collected by means of questionnaires completed by the parents. Eight subjects who performed CA challenge had no subsequent information available and were thus excluded. Since very few subjects had a follow-up longer than 6 yr, follow-up was truncated 6.0 yr after the CA challenge. The mean follow-up period after CA challenge was 4.4 ± 1.1 yr.

This study was approved by the Human Subjects Committee at the University of Arizona, and informed consent was obtained from the parents.

Definitions

Wheezing LRIs during the first 3 yr of life were assessed using the pediatrician's reports. Because some children changed pediatrician, not all subjects had information about wheezing LRIs during the first 3 yr of life despite their continued participation in the study, resulting in a smaller number of subjects for analyses when the data were stratified for wheezing LRIs (see RESULTS).

For the other respiratory signs and symptoms, questionnaires completed by parents were used. Parental history of asthma was defined as a history of asthma in either parent as reported in the enrollment questionnaire. Persistent cough was defined as at least two episodes of cough without a cold during the previous year. Wheezing was assessed by the question "Has this child's chest ever sounded wheezy or whistling?" Information about the number of episodes of wheezing during the previous year was obtained with the question "During the past year, how often has this child had wheezing or whistling?" The questions "Does this child cough during or shortly after vigorous exercise?" and "Does this child's chest sound wheezy or whistling during or shortly after vigorous exercise?" were used at age 6 to assess cough and wheezing after exercise, respectively. A child was considered to have developed asthma if he/she had, at the age 8 or age 11 survey, either: (1) asthma diagnosed by a physician with at least one episode of asthma or wheezing during the previous year or (2) more than three episodes of wheezing during the previous year regardless of a diagnosis of asthma.

Skin Tests

Skin prick tests were performed at the age 6 survey and included the following aeroallergens (Hollister-Stier Laboratories, Everett, WA): house-dust mix, Bermuda grass, careless weed, olive, mesquite, mulberry, and alternaria. Histamine and a negative control consisting of 50% glycerin were also applied. A child was considered to be atopic if he/she had a positive skin test reaction to at least one aeroallergen (> 3 mm after subtraction of the negative control value).

Cold Air Challenge

The mean ± SD age at the time of the CA challenge was 6.1 ± 0.5 yr. Children who were wheezing or needed bronchodilators at the time of the CA challenge, or who had LRIs during the previous 6 wk or upper respiratory tract infections during the previous 3 wk, were temporarily excluded from the challenge and rescheduled for testing.

The equipment used for CA generation is described in detail elsewhere (11). Cold dry air (15 to 25° C, < 1% relative humidity) was blown through a circuit connected to a mouthpiece by a Collins two-way J valve (Warren E. Collins, Inc., Braintree, MA). CO₂ was measured at the expiratory port with a CO₂ analyzer (Beckman, Fullerton, CA) and added into the circuit to keep the CO₂ concentration of the exhaled air and bias flow at 5%.

Lung function was assessed by asking the subjects to perform voluntary partial expiratory flow-volume curves. Tidal flow-volume curves were displayed on a computer monitor. At end-tidal inspiration, children were encouraged to exhale air forcefully, and the maximal expiratory flow at functional residual capacity (maxFRC in milliliters per second) was recorded.

Baseline lung function was defined as the best maxFRC among at least three acceptable prechallenge expirations. Subjects were then asked to breathe the CO₂-enriched CA for 6 min. An arcade-video game, driven by two No. 2 Fleisch pneumotachs, was used to encourage the children to keep a minute ventilation of 18 to 20 L/min. The mean of the first two values within the first 5 min after the challenge was taken as postchallenge value. Percent fall in maxFRC after CA challenge was calculated using the following equation:

A subject was defined as having bronchial hyperresponsiveness to CA challenge (CABHR) if his/her percent fall in maxFRC after CA was larger than the ninetieth percentile of the fall for the reference subjects (skin test-negative, never-wheezing, never asthma-diagnosed children as reported in the questionnaires at age 6). The cutoff value defined in this way corresponded to a percent fall of 41.1%.

Statistical Analysis

Unpaired Student t test and chi-squared test were used for the comparison of means and proportions. The 95% confidence intervals (95% CI) for relative risk (RR) were calculated with standard algorithms. Two-sided  levels < 0.05 were considered to be significant. To assess the incidence of asthma during the 6-yr follow-up, unadjusted Kaplan Meier failure curves were used, and comparisons of incidence rates were determined using the log-rank test (12). Hazard ratios (the likelihood in time of developing asthma for a child with a given risk factor divided by the likelihood in time of developing asthma for a child without that same risk factor) were calculated using the Cox proportional hazards model (12). Cox models were also used for multivariate analysis and were constructed by selecting variables found to be significantly associated with the dependent variable in the univariate analysis, or reported to be related to asthma in the literature (13). Variables considered as potential predictors of subsequent asthma were gender, CABHR, current mild wheezing (one to three episodes during the 12 mo prior to the age 6 questionnaire), skin test reactivity to allergens, maternal education ( or > 12 yr), current maternal smoking at age 6, and parental history of asthma. Because children were tested at different ages, the age at which the CA challenge was performed was included in the multivariate models.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A total of 1,024 children had questionnaires at the age 6 survey. Table 1 compares the 368 subjects who had not developed asthma by age 6 and who performed CA challenge to the rest of the population not included in this analysis (n = 656). Male gender, current maternal smoking, wheezing LRIs before age 3, persistent cough, atopy, and atopy and mild wheezing (one to three episodes of wheezing during the previous year) together were significantly lower in the study group than in the rest of the population. However, after stratifying for wheezing LRIs during the first 3 yr of life (Table 2), most of the differences between the study group and the rest of the population were still present only among children who had wheezing LRIs. Among children who did not have wheezing LRIs, subjects in the study group were not significantly different from the rest of the population, except for the lower proportion of children with both atopy and mild wheezing in the study group than in the rest of the population (8.5% versus 17.6%; p < 0.01). The groups did not differ in age at which CA challenge was performed. Children who had wheezing LRIs before age 3 had 3.2 times (95% CI = 2.4-4.3; p < 0.001) more asthma (as defined above) by age 6 than children who did not have wheezing LRIs.

Children who had mild wheezing at age 6 (one to three episodes of wheezing during the previous year) were more likely to be atopic and to cough or wheeze after exercise at age 6 than children who did not have mild wheezing (atopy: 47.1% [32 of 68] versus 33.2% [97 of 292], p < 0.05; cough after exercise: 32.4% [22 of 68] versus 13.0% [39 of 299], p < 0.001; wheezing after exercise: 14.5% [10 of 69] versus 1.7% [5 of 299], p < 0.001).

At age 6, children who were CABHR positive were more likely to have mild wheezing, atopy, and atopy and mild wheezing together than children who were CABHR negative (mild wheezing: 38.5% [20 of 52] versus 15.5% [49 of 316], p < 0.001; atopy: 59.6% [31 of 52] versus 31.8% [98 of 308], p < 0.001; atopy and mild wheezing together: 26.9% [14 of 52] versus 5.8% [18 of 308], p < 0.001). In Table 3, the proportions of CABHR-positive children among subjects with potential risk factors for BHR are compared with those without risk factors. Only mild wheezing (RR = 2.7, 95% CI = 1.7-4.4), atopy (RR = 2.6, 95% CI = 1.6-4.4), and wheezing and atopy together (RR = 3.8, 95% CI = 2.3-6.2) were significantly associated with CABHR among children who had not developed asthma by age 6. 

Of the 360 children who performed CA challenge and who had information about respiratory symptoms after age 6, 35 developed asthma (as defined above) by age 11; 30 were given a diagnosis of asthma by a physician and five developed recurrent wheezing (more than three episodes during the year prior to the questionnaire) without a physician diagnosis of asthma. The cumulative incidence of asthma between ages 6 and 11 (as assessed by Kaplan Meier failure curves) was 12.0%. Children who were CABHR positive at age 6 were more likely to develop new asthma between ages 6 and 11 than children who were CABHR negative (hazard ratio = 2.6, 95% CI = 1.2-5.4; p = 0.01) (Figure 1). Male gender (hazard ratio = 2.3, 95% CI = 1.2-4.5; p = 0.02), current mild wheezing (one to three episodes during the previous 12 mo) at age 6 (hazard ratio = 7.5, 95% CI = 3.8-14.7; p < 0.001), and atopy at age 6 (hazard ratio = 3.7, 95% CI = 1.8-7.4; p < 0.001) were also significantly associated with onset of asthma in the univariate analysis. Children with parental history of asthma were almost twice as likely to develop asthma after age 6 (hazard ratio = 1.9, 95% CI = 0.9-4.0; p = 0.07) when compared with children with no such history. Age at the time of CA challenge, maternal education, and maternal smoking at age 6 were not significantly associated with subsequent asthma (data not shown).

View larger version (12K): [in this window] [in a new window]   Figure 1.   Cumulative incidence of asthma after stratifying for CABHR in the whole study group (n = 360).

Multivariate analysis showed that, after adjusting for current mild wheezing at age 6, skin test reactivity to allergens at age 6, and other covariates, only mild wheezing and skin test reactivity to allergens were significant and independent predictors of subsequent onset of asthma, whereas CABHR was no longer significantly associated with new cases of asthma after age 6 (Table 4). In fact, subjects without atopy or current wheezing at age 6 who were CABHR positive at that age were no more likely to develop subsequent asthma than nonwheezing nonatopic subjects who were CABHR negative (1 of 15 [6.7%] versus 7 of 176 [4.0%]; p = 0.6). No association was found between wheezing LRIs before age 3 and asthma by age 11 among nonasthmatic nonwheezing nonatopic children at age 6. 

Having a wheezing LRI during the first 3 yr of life was not a significant predictor of the development of new cases of asthma after age 6 either univariately or after adjusting for CABHR (adjusted hazard ratio = 1.4, 95% CI = 0.7-3.0; p = 0.35). After stratifying by history of wheezing LRIs during the first 3 yr of life, CABHR was a strong predictor of asthma only among children who did not have wheezing LRIs (hazard ratio = 3.8, 95% CI = 1.6-9.2; p = 0.001) (Figure 2, upper panel). Among children who had wheezing LRIs during the first 3 yr of life, CABHR was not significantly associated with a higher incidence of asthma between ages 6 and 11 (hazard ratio = 1.6, 95% CI = 0.3-7.5; p = 0.6) (Figure 2, lower panel).

View larger version (12K): [in this window] [in a new window]   View larger version (12K): [in this window] [in a new window]   Figure 2.   (Upper panel ) Cumulative incidence of asthma after stratifying for CABHR among children who did not have wheezing LRIs before age 3 (n = 240). (Lower panel ) Cumulative incidence of asthma after stratifying for CABHR among children who had wheezing LRIs before age 3 (n = 80).

Multivariate analysis confirmed that, in children who did not have wheezing LRIs before age 3, only atopy and mild wheezing at age 6 were significantly associated with the development of new asthma after adjusting for potential confounders (Table 5). Children who were CABHR positive were 2.2 times more likely to develop asthma than CABHR-negative children after adjusting for the other covariates, but this association did not reach statistical significance (95% CI = 0.8-6.2; p = 0.121). In children who had wheezing LRIs before age 3, atopy and mild wheezing at age 6 were also the only two significant predictors of new cases of asthma after age 6 (atopy: adjusted hazard ratio = 18.0, 95% CI = 1.1-300.0, p = 0.04; mild wheezing: adjusted hazard ratio = 11.7, 95% CI = 1.9- 72.1, p < 0.01) after adjusting for the same covariates shown in Table 5. No interaction between the different risk factors was found to be significant either among children who did not have wheezing LRIs before age 3 or among those who had wheezing LRIs before age 3. 

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We found that CABHR assessed at a mean age of 6 yr was a significant predictor of subsequent diagnosis of asthma and/or severe asthma-like symptoms in children who did not have a diagnosis of asthma at the time of the CA challenge. However, after adjusting for potential confounders, this effect was no longer significant, and atopy and current mild wheezing (one to three episodes) at age 6 were the only significant and independent predictors of subsequent asthma. In children who did not have wheezing LRIs before age 3, CABHR was associated with a 2.2-fold risk of developing new asthma after adjusting for the other covariates, but this association did not reach statistical significance.

Many of the studies on BHR and subsequent development of asthma in children have been performed using either methacholine challenge or histamine challenge. Although it has been pointed out that in adults suspected of having asthma a positive methacholine challenge not always corresponds to a positive CA challenge and vice versa (14), most of the studies in the literature show a strong correlation between the response to CA and the response to either histamine (15) or methacholine (15). It has been suggested that CABHR may be more specific for asthma than methacholine BHR (19).

Several studies have suggested that BHR can predict the development of new asthma in children. Hopp and coworkers (4) studied 13 subjects before and after the onset of clinical asthma; the mean age at initial visit was 10.6 yr and the mean follow-up interval was 6.3 yr. They found that 10 of the 13 subjects had moderate to marked bronchial hyperreactivity at the initial visit as assessed by methacholine challenge. This was significantly different from the response to methacholine of 13 subjects from nonasthmatic families who did not develop asthma but not significantly different from the response to methacholine of 13 subjects from asthmatic families who did not develop asthma. However, five of the 13 study subjects had wheezing at the initial visit and the number of subjects studied was very small. Peat and colleagues (5) followed 236 subjects 8 to 11 yr of age over a period of 1 yr. BHR was assessed by means of methacholine challenge and information about respiratory symptoms was collected using questionnaires completed by parents. At the beginning of the study, children were categorized into four groups: "current asthma" (BHR + wheeze), "wheeze only," "BHR only," and "normal." By the end of the study, the "BHR only" group had a prevalence of wheeze of 52% versus 18% in the "normal" group (p < 0.001). However, atopy was significantly higher in the "BHR only" group than in the control group, and no adjustment for atopy was made. Jones (7) studied 55 asymptomatic BHR-positive children compared with 55 asymptomatic BHR-negative control subjects (age range: 5 to 11 yr). BHR was assessed using a free running exercise test. At the follow-up 6 yr later, 58% of the subjects in the BHR group had developed asthma versus 13% of the subjects in the control group (p < 0.001). Once again, however, the BHR group was not truly asymptomatic at the beginning of the study, since prevalence of cough, wheezing, and bronchitis was significantly higher in this group than in the control group. No adjustment was made for skin test reactivity to allergens. Carey and associates (8), who studied 281 adolescents over a period of 18 mo, found that BHR, assessed by means of CA or methacholine challenge, was significantly and independently associated with a greater likelihood of developing wheezing by the end of the study. Only subjects with prevalent wheeze at the time of testing were excluded from the analyses. Conversely, de Gooijer and coworkers (6) reported that only atopy (but not BHR as assessed by histamine challenge at ages, 8 to 11 yr) was a significant predictor of onset of asthma-like symptoms in young adulthood. It is important to point out that, in all these studies, information about the years prior to the investigation was either collected retrospectively, which could have introduced preferential recall bias, or limited to the previous 12 mo, thereby failing to account for earlier symptoms or diagnoses.

In several of the previous reports, no adjustment was made for mild asthma-like symptoms at the time of BHR testing in subjects who subsequently developed asthma. In our study, mild wheezing (one to three episodes) at age 6 was the strongest risk factor for developing a diagnosis of asthma or asthma-like symptoms between ages 6 and 11. It is possible that, in children with BHR, infrequent wheezing may already represent a mild form of asthma (5, 20). A recent report suggests that BHR in asthmatic children is associated with evidence of airway inflammation (21). The underlying pathophysiology of BHR in asymptomatic children, however, remains to be elucidated. In asymptomatic subjects, BHR might represent a manifestation of subclinical inflammation of the airways (22). Pin and coworkers (19), however, found that asymptomatic BHR-positive children had eosinophil counts in the sputum that were not significantly different from those in normal control subjects and were lower than those in asthmatic BHR-positive children. In addition, BHR in asymptomatic subjects did not improve after treatment with inhaled budesonide. These findings suggest that symptomatic and asymptomatic BHR-positive children may have different inflammatory patterns in the airways.

Atopy was also a strong predictor of onset of asthma between ages 6 and 11. This finding is in agreement with the above-mentioned study by de Gooijer and coworkers (6) and is also consistent with the data by Sears and colleagues (23), who found a strong positive association between the prevalence of BHR and the level of total serum IgE as a measure of atopy in 11-yr-old children. Recently, this same group (24) reported that there is also a close association between BHR and the size of the reaction to a panel of allergy skin tests both in asthmatic and nonasthmatic subjects. Moreover, twin studies (25, 26) suggest that BHR may be mainly an acquired phenomenon secondary to atopy. In addition, Postma and associates (27) found a strong association between BHR and atopy and suggested that BHR may be co-inherited with an elevated level of serum total IgE.

We found that BHR assessed at a mean age of 6 by CA challenge was a significant univariate predictor of new asthma by age 11 among children who did not have wheezing LRIs during the first 3 yr of life and who did not have asthma at age 6. We did not find such an association among children who had wheezing LRIs before age 3. The most likely explanation for this discrepancy is that children who wheeze during the first 3 yr of life tend to develop a diagnosis of asthma by age 6 more often than those who do not wheeze during the first 3 yr of life (28). Because we excluded from the analysis all subjects with asthma and/or severe or frequent wheezing by age 6, most children from the wheezing LRI group who were potentially CABHR positive were probably excluded from this analysis because they had already developed asthma or severe wheezing before age 6.

It has been suggested that bronchiolitis could cause subsequent BHR (29) and that BHR in asymptomatic children may be due to previous insults (19), such as wheezing LRIs during the first years of life. In our data, there was no association between wheezing LRIs before age 3 and CABHR at age 6. However, among CABHR-negative children, the cumulative incidence of asthma was higher (albeit not significantly) in children who had a history of wheezing LRIs than in those who did not have a history of wheezing LRIs (11.6% versus 6.8%, p = 0.178). We thus cannot exclude the possibility that wheezing LRIs during the first 3 yr of life may be associated with the development of asthma after age 6 independently of BHR.

In children who did not have wheezing LRIs before age 3, CABHR was associated with a 2.2-fold increased risk of developing new asthma after adjusting for the other covariates, but this association did not reach statistical significance. A possible explanation for this finding is that, in children who did not have wheezing LRIs, CABHR at age 6 might actually be an independent predictor of subsequent asthma, but this effect may perhaps not be strong enough to be statistically significant in our data set after adjusting for mild wheezing at age 6, skin test reactivity to allergens, and other covariates. A genome-wide search for BHR-linked loci by Daniels and colleagues (30) found that only one out of three loci showing evidence of linkage to BHR did not co-segregate with markers for atopy. In other words, BHR might be mostly co-inherited with atopy (27), but in some cases (perhaps a minority) BHR might be independent of atopy. It has also been suggested that the association between BHR and atopy might simply be the result of the high prevalence in the population of two independently heritable traits (31). A preliminary report by Gibson and associates (32) also suggests that BHR, as assessed by histamine challenge at the age of 1 mo, predicts asthma, atopy, and level of BHR at the age of 6 yr, implying that the level of BHR may already be partially set from birth. Clearly, the relationship between BHR, atopy, and respiratory symptoms is still not completely understood (33). More studies on the genetics and ontogeny of BHR and of the immune response system of the lung will certainly contribute to elucidate this complex issue.

Finally, it needs to be noted that, given the high variability of BHR measurements in time, a single BHR test may not give reliable information about the usual responsiveness of the airways and the tendency to subsequently develop respiratory symptoms. A recent report by Burrows and coworkers (34) shows that repeated positive methacholine tests between age 9 and age 15 predicts almost 85% of current wheezing at age 15.

In summary, CABHR measured as early as at a mean age of 6 yr was a significant univariate predictor of asthma and/or asthma-like symptoms in children who did not have a diagnosis of asthma by age 6. However, after adjusting for other covariates, this association was no longer statistically significant, and atopy and current mild wheezing at age 6 were the only significant and independent predictors of subsequent asthma.