The natural history of asthma is still poorly understood [
This retrospective study was performed in two tertiary care centres: the Asthma and Allergies Centre, Armand-Trousseau Hospital, Paris (France), and the Centre for Airborne Allergies in Infants, CHU-Estaing (University Teaching Hospital), Clermont-Ferrand (France).
We recruited infants aged under 36 months who had a history of at least three wheezing episodes and who had been assessed for respiratory wheezing disease via a doctor-led ISAAC questionnaire [
The patients were contacted at the age of 13 years by a doctor, who used the ISAAC questionnaire in a telephone interview to assess the pattern of active asthma symptoms over the preceding 12 months.
The predictive parameters measured and collated were (i) demographic characteristics: age, gender; (ii) parental documented history, when available, of doctor-diagnosed asthma and infant’s personal history of atopic dermatitis; (iii) severe recurrent wheezing indicated by at least two previous hospital admissions due to exacerbation; (iv) biological markers of atopy: eosinophilia (hypereosinophilia was defined as an absolute eosinophil count of ≥470/mm3) [
Persistent active asthma in the previous 12 months was defined as the presence of asthma symptoms and/or antiasthma medication taken during that period.
The study was performed in accordance with French and international best practices for epidemiological research. Consent forms were collected from the parents for every questionnaire. The Saint-Antoine Hospital (Paris) institutional review board for biomedical research did not require approval from an ethics committee, since the study protocol did not feature any investigations other than those routinely used in the day-to-day management of the patient. Data analysis was managed in such a way as to protect confidentiality as stipulated in the French
Statistical analysis was performed using STATA v11 for Windows (StataCorp, Tex, USA). Continuous quantitative variables were expressed as means ± standard deviation and qualitative variables as numbers and percentages. The dependent variable studied was active asthma symptoms at age of 13.
Under univariate analysis, continuous variables from the asthmatic
Predictive variables for persistence of asthma at age 13 after univariate analysis (
The two centres recruited a combined total of 541 patients, for 436 of whom fully workable datasets were compiled. We were able to recontact 227 of the 436 at age 13, who then formed the final study population. Of the remaining patients, 206 were registered as lost-to-followup (no valid address at call-up or three or four calls left without a reply) and 3 responded to call-up but declined to complete the questionnaire. The age of children lost-to-followup was significatively lower at initial diagnosis than the average age of recruits (
Average age at the intake assessment was 26 months (SD: 8.8). Boys accounted for 68.1% of cases (
The factors significantly associated with active asthma at the follow-up end point were presence of initial atopic dermatitis (
Predictive parameters of asthma persistence at follow-up end point (univariate analysis).
Persistent | Remission | ||
---|---|---|---|
Atopic dermatitis | 85 (70.3%) | 39 (36.8%) | <0.001 |
Sensitization to ≥2 food allergens | 15 (12.4%) | 4 (3.7%) | 0.02 |
Sensitization to ≥2 airborne allergens | 40 (33.1%) | 7 (6.6%) | <0.001 |
Allergen polysensitization | 47 (38.8%) | 12 (11.3%) | <0.001 |
Total serum IgE >45 IU/mL | 59 (48.8%) | 30 (28.3%) | 0.002 |
Blood eosinophil count ≥470/mm3 | 39 (32.2%) | 15 (14.2%) | 0.001 |
Mother with asthma | 27 (22.5%) | 20 (19.1%) | 0.52 |
Father with asthma | 28 (23.3%) | 19 (18.6%) | 0.33 |
Initial severity | 62 (51.2%) | 34 (32.1%) | 0.004 |
Allergen sensitization: presence of allergen-specific IgE ≥0.35 kU/L to at least one of the common aeroallergens (dust mites, cat or dog dander, seed plant, or birch pollen) or one of the common food allergens (cow’s milk, eggs, peanuts, wheat, soybean, and fish). Allergen sensitization was single or multiple (≥2 specific food allergen-positive IgE or ≥2 specific airborne allergen-positive IgE). Allergen polysensitization was defined as ≥2 allergen-specific IgE, regardless of allergen class. Initial severity: two previous hospital admissions due to exacerbation. Hypereosinophilia: absolute eosinophil count of ≥470/ mm3.
The most relevant risk factors for persistent asthma were: sensitization to multiple aeroallergen (OR 4.6, CI-95% (1.9–11.2) previous history of atopic dermatitis (OR 3.4, CI-95% (1.9–6.3) initial severity (OR 2.3, CI-95% (1.3–4.2) hypereosinophilia ≥470/mm3 (OR 2.2, CI-95% (1.07–4.7)
Finally, of the asthmatic infants presenting all these risk factors, 95.6% remained active asthmatics in adolescence, with 81% of the cases being classifiable as mild to severe persistent.
Sensitivity, specificity, negative and positive negative values, and ROC Curve Area are given in Table
Assessment of the ability of modified API to predict time course development of asthma in adolescence in the study population.
Asthma at the follow-up end point | Sensitivity % (CI-95%) | Specificity % (CI-95%) | PPV %(CI-95%) | NPV %(CI-95%) | Area under the ROC curve | Correctly predicted, % |
Positive API | 87 (79–92) | 37 (28–47) | 61 (53–68) | 71 (57–82) | 0.62 | 68.7 |
PPV: positive predictive value; NPV: negative predictive value.
ROC: receiver operating characteristic curve.
API: asthma predictive index.
The main advantage of this study is that it was performed bicentrically, with cohorts of patients submitted to the same management procedure for both the questionnaire and the allergy testing programme. However, our findings have some limitations because of the retrospective nature of the study and the percentage of cases lost to followup (52.2%). This percentage is high but needs to be seen in the perspective of the length of the follow-up period, which ran to over 10 years. Other long-term follow-up studies have reported similar findings [
Unfortunately, we have no data on respiratory function parameters to corroborate persistently active asthma in the previous 12 months [
At the age of 13, 121 patients (55.8%) have had active asthma over the previous 12-month period. In an earlier study of the Paris-based cohort, asthma symptoms were observed in only
The risk factors associated with persistent childhood asthma symptoms were consistent with those in the literature: multiple sensitization to aeroallergens in recurrent wheezing infants as a major risk factor for active asthma in adolescence [
Several authors had tried to establish scores to predict the long-term persistence of asthma. However, these scores do not have high statistical predictive power, especially when used in cohorts of wheezing infants instead of general-population infants. The most widely used score, the modified asthma predictive index (API), offers the benefit of being easy to use, but sensitivity, specificity, and positive predictive value are quite low [
Predicting the long-term course of asthma is still difficult, although atopy is confirmed as a major predictive risk factor. Predictive scores are unfortunately not reliable enough performances to have a real value at an individual level. Early identification of clinical and biological signs of atopy is essential to target the population of wheezing infants who will need specialized long-term followup and who would greatly benefit from specific treatment.
The authors declare that there is no conflict of interests.
The authors would like to thank Mr. Jeffrey Watts for his careful reading.