DISCUSSION
Taking together, these results identified independent factors associated with subsequent exacerbation frequency rates within the year following fibreoptic bronchoscopy in severe preschool wheezers. These factors include atopic dermatitis, age at first episode of wheezing, the presence of Haemophilus in the BALF, the RBM-BSM distance and the BSM area. The BSM area was the most important factor and was associated with each exacerbation frequency group.
The risk of subsequent exacerbations has been previously associated with a range of various parameters such as markers of uncontrolled disease (i.e. , uncontrolled symptoms of wheezing, at least one severe exacerbation in the previous year) in preschool wheezers (12,17). Thus, it was not surprising to note that three quarters of the patients in our study that included severe preschool children who experienced at least one exacerbation or one hospitalization for wheeze within a year prior to inclusion, continued to experience exacerbations during the follow-up year. Moreover, using univariate multinominal regression analysis, we identified a significant association between the number of hospitalizations within a year prior to inclusion and the exacerbation frequency in the follow-up year in agreement with previous reports (12,17).
The risk of exacerbations in preschool wheezers has also been associated with type-2 inflammation, highlighted by the presence of atopic dermatitis or allergic sensitizations, increased levels of blood eosinophils and/or exhaled NO levels (17–21). On the one hand, we do confirm, in the present study, that atopic dermatitis was an independent risk factor associated with the exacerbation frequency in severe preschool wheezers and that its prevalence was increased in the high exacerbation group in agreement with previous studies (17–19,21). However, the presence of atopic dermatitis was also associated with the no-exacerbation group. Therefore, this result suggest that isolated atopic dermatitis was not a risk factor of subsequent exacerbations but when combined with other exacerbation risk factors, its presence increased this risk. On the other hand, previous studies demonstrated an increased risk of exacerbations in atopic and/or smoking mothers’ infants with elevated exhaled NO levels after birth (20) and in preschool children with single (17) or early multiple allergic sensitizations (18,21) or increased blood eosinophils (19). However, we did not find any difference in exacerbation rate according to allergic sensitization or blood eosinophils levels. The association between exacerbation risk and blood eosinophil levels or allergic sensitization has previously been reported in populations including patients treated or not with ICS (17–19,21). In our study, all patients were treated at inclusion with high doses of ICS with a second controller which may have decreased the impact of both blood eosinophils level and allergic sensitization on the risk of exacerbations (22). Indeed, Fitzpatricket al . showed that the risk of exacerbations was significantly decreased by ICS and abolished the statistical significance between the sensitized or blood eosinophils based groups (18,19).
Non-type 2 inflammation related parameters (i.e. , neutrophilic inflammation, isolation of bacteria on BALF culture) have also been associated with an increased risk of subsequent exacerbations (9,11). Despite no significant differences between exacerbation frequency-groups according to BALF and bronchial inflammation, the univariate analysis showed that a decreased percentage of macrophages and an increased percentage of PMN were significantly associated with infrequent exacerbations whereas a low density of lymphocytes in bronchial biopsies was a significant risk factor for frequent exacerbations. Unfortunately, these parameters were not included in the multivariate analysis due to the great number of missing values. We also demonstrated that bacteria in the BALF represent an independent risk factor for subsequent exacerbations but was mostly associated with a low annual exacerbation rate. This was consistent with studies reporting the efficacy of antibiotic or bacterial lysate therapies to decrease wheezing attacks rates and the use of oral corticosteroids compared to placebo (11,23,24).
In the present study, the risk of subsequent exacerbations was associated with several bronchial remodelling parameters (i.e ., submucosal fibrosis area, BSM area, RBM thickness, RBM-BSM distance) as previously reported (25–28). However, only the BSM area and the RBM-BSM distance appeared as independent risk factors of subsequent exacerbations in severe preschool wheezers using multivariate analysis. Moreover, it is interesting to note that the group with the highest exacerbation frequency is characterized by the greatest BSM area, the lowest RBM-BSM distance and the earliest age at first episode of wheezing. Thus, it is tempting to speculate that early-life respiratory tract infections promote BSM remodelling which in turn increased the risk of virus- and/or bacteria-induced exacerbation creating a vicious circle. Indeed, the first year of life appeared as a period of high susceptibility of BSM remodelling (29) characterized by an increased mitochondria-dependent BSM cell proliferation (30), which has been involved in BSM remodelling of both preschool wheezers (31) and adult asthma (32). Moreover, in adult asthma, we recently demonstrated an increased viral replication within the bronchial epithelium induced by the BSM which can increased exacerbation susceptibility (33) and in turn induced a specific BSM cell migration toward bronchial epithelium decreasing RBM-BSM distance (34). However, further mechanistic studies would be necessary to assess both the underlying mechanisms of increased BSM mass and the impact of BSM remodelling on viral replication in the specific population of preschool wheezers.
All these findings reinforce the usefulness of fibreoptic bronchoscopy in the specific population of severe preschool wheezers. Indeed, while it can be argued that fibreoptic bronchoscopy with bronchial biopsies is an invasive procedure requiring anaesthesia and of limited interest especially in such young children. Our study has, on the contrary, demonstrated its usefulness to provide information on bronchial inflammation, dysbiosis and remodelling and then on the risk of subsequent exacerbations. This well-tolerated examination allows for better characterization of the disease and helps the clinician to optimize the therapeutic strategy with a favourable benefit-risk ratio in severe preschool wheezing (14).
Several limits of the present study must be pointed out: (i ) The relatively small number of patients limited the size of each group may have increased the risk of type 2 error in the group comparisons and the 95% confidences interval in the logistic regression analysis. In addition, the small size of our population has limited the maximal number of parameters included in the multinomial logistic regression model number. Indeed, some of the significant parameters in univariate analyses (i.e. , Macrophages and PMN in the BALF as well as density of lymphocytes) had too many missing values in bronchial biopsies, and were assessed only in the P’tit Asthme” cohort (14). However, this pooled cohort represent the largest cohort describing both clinical and fibreoptic bronchoscopy parameters in severe preschool wheezers. (ii ) As highlighted recently (35), our data did not include long term follow-up data (i.e ., more than one year) allowing for the assessment of the stability of exacerbation frequency over time and the consequences of bronchial remodelling on lung function and asthma persistence at school age. Additional, long term longitudinal studies are thus necessary.