INTRODUCTION — Parents or caregivers of infants and young children with recurrent wheezing often ask, "Does my child have asthma?" This is a question that clinicians involved in patient care and also those researching asthma would like to be able to answer. Preschool wheezing, a symptom that can herald the subsequent development of childhood asthma, is a common problem worldwide [1]. However, the condition improves and ultimately disappears by school years in many children. Proper identification of infants and young children at increased risk to develop persistent asthma may help predict long-term outcomes and improve prevention and treatment, but the ability to identify these children remains limited.
Several classifications of wheezing phenotypes and other tools have been developed in an effort to categorize children with recurrent wheezing and determine which will ultimately develop asthma. These wheezing phenotypes and predictive tools are reviewed in this topic. The definition and diagnosis of asthma in children; asthma risk factors, genetics, and natural history; bronchiolitis and virus-induced wheezing; and other causes of wheezing in children are discussed separately. (See "Asthma in children younger than 12 years: Initial evaluation and diagnosis" and "Natural history of asthma" and "Genetics of asthma" and "Risk factors for asthma" and "Evaluation of wheezing in infants and children" and "Role of viruses in wheezing and asthma: An overview" and "Treatment of recurrent virus-induced wheezing in young children" and "Bronchiolitis in infants and children: Clinical features and diagnosis".)
WHEEZING PHENOTYPES — Almost 50 percent of children are reported to have wheezing in the first year of life, although only 20 percent will experience continued wheezing symptoms in later childhood [2]. Wheezing phenotypes have been defined to identify the characteristics and risk factors associated with children that experience wheezing [3]. Some of these phenotypes describe children who continue to wheeze until later childhood, while others identify those who continue to wheeze through adolescence and adulthood. However, the relationship between risk factors and the subsequent development of asthma in later childhood and adult life is not clear.
Many of these early-childhood wheezing phenotypes were determined retrospectively in longitudinal studies. However, it can be difficult to clinically distinguish among these phenotypes during the preschool years because of the variation in expression of both symptoms and risk factors over time. In addition, the application of these phenotypes to diverse populations is not established, and it is not clear which are the most effective therapies for a particular wheezing phenotype or whether early intervention can alter the course and outcome over time.
Epidemiologic phenotypes — Several epidemiologic phenotypes, based upon wheezing history, have been developed. There is some variability in age cutoffs differentiating among transient, persistent, and late-onset wheezing and inconsistencies in which risk factors are associated with each phenotype. There are also concerns about the prospective validity of these phenotypes.
Tucson classification — In the Tucson Children's Respiratory Study (TCRS), four different wheezing phenotypes were identified among 1246 newborns followed for lower respiratory tract infections based upon the presence of wheezing symptoms during the first three years of life and again at six years [2]. The epidemiologic phenotypes generated from this prospective longitudinal study included:
●Never wheezers (51 percent) – Healthy children who never wheezed.
●Early, transient wheezers (20 percent) – Children with wheezing that began before three years of age and resolved by six years of age.
●Persistent wheezers (14 percent) – Children with wheezing that began before three years of age and was still present at six years of age.
●Late-onset wheezers (15 percent) – Children who developed wheezing between three and six years of age.
Children in the persistent wheezing and late-onset wheezing groups are at increased risk for persistent asthma-like symptoms into adolescence and adulthood.
A report from the Italian Studies of Respiratory Disorders in Childhood and the Environment (SIDRIA) found a different frequency of the above phenotypes [4]. The classification was the same as the Tucson classification with the exception that the cutoff for distinguishing between transient, early wheezing and persistent wheezing was two, not three, years of age. Of the 16,333 children aged six to seven years in this population-based cohort, 83 percent had never wheezed; 7 percent had transient, early wheezing; 4 percent had persistent wheezing; and 6 percent had late-onset wheezing.
Subsequent prospective studies of the Tucson cohort led to revised definitions of the three groups of wheezers [5-7]:
●Transient wheeze in infancy – Begins in infancy (the first year of life) and resolves by the preschool years; associated with decreased lung function, narrower intrapulmonary airways, maternal tobacco use during pregnancy, having siblings, and daycare attendance.
●Nonatopic persistent wheezing phenotype – Begins in infancy and resolves in mid-childhood; associated with lack of both allergic sensitization and methacholine hyperresponsiveness.
●Immunoglobulin E (IgE) associated/atopic, persistent wheezing phenotype – Can begin in infancy but increases in prevalence with age; associated with personal and family history of atopy, methacholine hyperresponsiveness, and poor growth of lung function. This phenotype may represent a classic allergic asthma phenotype, but it is unknown if children with this phenotype will have symptoms that persist into adulthood.
Avon classification — The Avon Longitudinal Study of Parents and Children (ALSPAC), a prospective, longitudinal study of a population-based cohort in the United Kingdom, looked at maternal reports of child wheezing between birth and six months and again between 30 and 42 months in 8594 children [8]. Early infant wheeze was defined as those who wheezed during the first six months, late-onset wheeze as those who wheezed between 30 to 42 months but not early infancy, and persistent wheeze as those who wheezed during both periods. A parental history of asthma (particularly maternal) and a personal history of atopic disease were predictive of wheezing in all three groups. Wheezing only during early infancy was associated with other factors, such as maternal smoking during pregnancy and the presence of older siblings, whereas persistent wheezing was associated with preterm delivery and lower socioeconomic status.
The wheezing phenotypes from the Avon study were further refined with continued longitudinal follow-up [9]. Data were collected at more time points and over shorter intervals than the Tucson study. Incorporation of objective measures (eg, lung function and allergen sensitization) and use of more sophisticated statistical methods led to the inclusion of two additional phenotypes to the Tucson classification (prolonged, early wheeze and intermediate-onset wheeze):
●Never/infrequent wheeze (59 percent)
●Transient, early wheeze (16 percent) – Wheezing common from 6 to 18 months but rare to never after 42 months
●Prolonged, early wheeze (9 percent) – Wheezing common from 6 to 54 months but rare to never after 69 months
●Intermediate-onset wheeze (3 percent) – Wheezing rare to never from 6 to 18 months but common thereafter
●Late-onset wheeze (6 percent) – Infrequent wheezing from 6 to 42 months but common thereafter
●Persistent wheeze (7 percent) – Wheezing common from six months onward
Similar wheezing phenotypes were identified in the Prevention and Incidence of Asthma and Mite Allergy (PIAMA) birth cohort [10]. These phenotypes were also analyzed from early childhood into adolescence by pooling data of 7719 participants from five birth cohorts [11]. Compared with the never/infrequent wheeze, the other phenotypes had higher odds of asthma and lower lung function in adolescence. The persistent wheezing and late-onset wheezing groups exhibited the highest prevalence of current asthma in adolescence (adjusted odds ratio [OR] 48.31 and 36.39, respectively) and use of asthma medications (adjusted OR 42.45 and 24.51, respectively). Considerable within-class heterogeneity was noted at the individual level, particularly in the late-onset wheeze group and early-onset, preschool-remitting wheeze groups.
Multizentrische Allergiestudie (MAS) and Protection against Allergy: Study in Rural Environments (PASTURE) — This study identified benign, symptomatic, and severe atopy phenotypes using latent class analyses of two cohorts of healthy German infants and rural European children followed during the first six years of life [12]. Although the severe atopy phenotype was found in only 5 percent of children in these cohorts, it was associated with an asthma rate of 20 percent. A pronounced increase in seasonal aeroallergen sensitization was observed before three to four years of age, followed by the development of high serum IgE levels and then the onset of asthma symptoms. Higher serum IgE production was linked to impaired lung function and increased asthma risk in later life.
Mechanisms of the Development of ALLergy (MeDALL) — The MeDALL database collects data from over 44,000 participants from 14 European cohorts [13]. It found that polysensitization, but not monosensitization, to aeroallergens was linked to asthma symptom onset in early life and the presence of other allergic comorbid diseases (eg, rhinitis, eczema). Children with the polysensitization phenotype also had more persistent symptoms, more severe asthma, and increased total and specific IgE levels. The researchers were unable to develop a predictive algorithm for asthma risk in young children, postulating that viral infections may have confounded the predictive value of preschool asthma-like symptoms.
Canadian Asthma Primary Prevention Study (CAPPS) — CAPPS identified three wheezing phenotypes (low progressive, early transient, and early persistent) in a cohort of 545 infants followed from birth to 15 years of age who were considered high risk for asthma due to first-degree relatives with allergic disease and who underwent a primary prevention intervention [14]. The low, progressive group had low rates of wheeze in early life, but the wheezing frequency increased to 16 percent by age 15 years. The early-life intervention was avoidance of house dust mites, pet allergens, and environmental tobacco smoke exposure, as well as promotion of breastfeeding and use of a partially hydrolyzed formula if supplementing. These interventions were found to decrease the risk of persistent wheezing during school age only in the early, persistent wheeze phenotype. Risk factors associated with the early, persistent phenotype included wheezing during the second, but not first, year of life; male sex; and maternal asthma. Those in the early, transient or early, persistent groups were more likely to have allergic sensitization at 12 months than those in the low, progressive group.
Urban Environment and Childhood Asthma (URECA) — Five wheezing phenotypes (low wheeze/low atopy, low wheeze/high atopy, transient wheeze/low atopy, high wheeze/low atopy, high wheeze/high atopy) were identified in a high-risk, urban cohort of 442 children followed between birth and seven years of age using latent class mixed models [15]. Early-life environmental exposures were found to differentiate these phenotypes. The high-wheeze/high-atopy and high-wheeze/low-atopy phenotype groups were associated with decreased indoor allergen exposure and an increased frequency of asthma by age seven years. The high-wheeze/high-atopy group also demonstrated increased disease burden. Prenatal smoke exposure and maternal stress and depression were associated with the high-wheeze/low-atopy group. The high-wheeze/high-atopy phenotype was associated with low household dust microbial diversity. Early-life environmental exposures may modify risk for wheezing phenotypes.
The wheezing phenotypes from the URECA cohort were updated to include repeated measures of wheeze, allergen-specific IgE, and lung function data from birth until 10 years of age [16]. These were linked to early-life exposures and nasal epithelial gene expression at age 11 years. Six wheezing phenotypes were identified, three of which demonstrated little or no atopy and differed by patterns of wheeze, and three that were defined by significant atopy, different patterns of wheezing, and lung function:
●Low wheeze/low atopy (22 percent) – Minimal wheeze
●Transient wheeze/low atopy (17 percent) – Wheezing during early life that generally had resolved by three to four years of age
●Moderate wheeze/low atopy (12 percent) – High rates of wheezing during infancy that steadily diminished through age 10 years
●Low wheeze/high atopy (19 percent) – High rates of allergic sensitization, little or no wheezing
●Moderate wheeze/high atopy (17 percent) – High rates of allergic sensitization, wheezing during infancy that steadily diminished with time
●High wheeze/high atopy/low lung function (13 percent) – High rates of allergic sensitization, high rates of wheezing that persisted through age 10 years, and evidence of airway obstruction
The high wheeze/high atopy/low lung function group had the most respiratory morbidity. Children with this phenotype had low exposure to common allergens and elevated exposure to ergosterol in house dust. All high atopy groups demonstrated increased expression of a type-2 inflammation gene module in nasal epithelial samples. However, epithelium interleukin (IL) 13 response was linked with impaired lung function, and the high wheeze/high atopy/low lung function group was associated with mucin 5AC, oligomeric mucus/gel forming (MUC5AC) hypersecretion. Conversely, altered expression of modules of epithelial integrity, epithelial injury, and antioxidant pathways were associated with the medium wheeze/low atopy group.
Symptom-based phenotypes — Epidemiologic phenotypes are limited by their retrospectively defined patterns of duration of wheeze. This limitation led to the proposal of phenotypes based upon temporal patterns of wheeze that can be applied prospectively and may therefore be more useful when making treatment decisions.
The European Respiratory Society defined two symptom-based phenotypes [17]:
●Episodic (viral) wheeze – Wheezing during discrete time periods, with absence of wheeze between episodes; usually associated with viral respiratory tract infections
●Multi-trigger wheeze – Wheezing both during discrete exacerbations and between episodes; triggers include viruses, allergens, exercise, and cigarette smoke
Airway function, particularly conductive airways ventilation inhomogeneity, was lower in multiple trigger than episodic wheeze, suggesting that these are functionally different phenotypes [18]. However, in one study, over half of children classified into these two phenotypes based upon their wheezing history in the previous year switched to the other phenotype in the ensuing year [19], suggesting that these phenotypes are not stable over time.
Another study compared epidemiologic phenotype definitions identified by latent class analysis (similar to the Avon classification: persistent, late onset, intermediate, transient, and none/infrequent) with clinical phenotypes based upon patient histories from one to six years of age in an international, multicenter birth cohort [20]. Four clinically defined wheeze phenotypes (recurrent, unremitting wheeze; unremitting wheeze; asthma diagnosis; and frequent wheeze) were found to have a high sensitivity and specificity (displayed area under the receiver operating characteristics curve of 97, 87, 85, and 84 percent, respectively). These findings suggest that these clinical phenotypes may be useful asthma definitions for future epidemiological studies. The clinical phenotypes were characterized as follows:
●Asthma diagnosis – Clinician diagnosis of asthma ever or recurrent diagnoses of spastic, obstructive, or asthmatic bronchitis by age six years
●Multi-trigger wheeze – At least two common asthma triggers leading to wheeze between ages three to six years
●Unremitting wheeze – Having symptoms between wheezing episodes or wheeze without a cold at least once between ages one to six years
●Recurrent, unremitting wheeze – Having symptoms between wheezing episodes or wheeze without a cold reported for two or more years between ages one to six years
●Frequent wheeze – Wheeze on a monthly basis for at least one year between ages one to six years
●Episodic wheeze – Wheezing episodes associated with viral respiratory tract infections only between ages one to six years
Trousseau wheezing phenotypes — Three wheezing phenotypes were identified in a cross-sectional analysis of 551 children less than 36 months of age with recurrent wheezing enrolled in the Trousseau Asthma Program in France [21]. These categories share some similarities with both the epidemiologic and symptom-based phenotypes. It is not known whether these phenotypes are persistent over time. Factors examined in the cluster analysis included disease severity, wheezing triggers, family and personal history of atopy, inflammatory markers, and chest radiograph results. Most of the children had uncontrolled or partially controlled wheezing and were treated with inhaled glucocorticoids.
The three wheezing clusters identified were:
●Mild, episodic, viral wheeze (59.3 percent) – Most children in this group had mild disease, wheezing triggered by viral upper respiratory infections (URIs), and normal chest radiographs. Wheezing symptoms were controlled without the use of high-dose inhaled glucocorticoids.
●Nonatopic, uncontrolled wheeze (28.5 percent) – Children in this group typically had moderate-to-severe disease with abnormal chest radiographs, parents with asthma, and continued wheezing despite treatment with high doses of inhaled glucocorticoids. This cluster had the highest percentage of females.
●Atopic, multiple trigger wheeze (12.2 percent) – This group was characterized by multiple wheezing triggers (eg, URIs, exercise, cold air), higher percentage of males, eczema, increased laboratory markers of atopy (eg, elevated total and specific IgE), and abnormal chest radiographs.
Project Viva wheezing phenotypes — Four longitudinal wheeze trajectories (never/infrequent wheeze; mid-childhood-onset wheeze; early, transient wheeze; and persistent wheeze) were identified in a Massachusetts-based prebirth, population-based cohort study [22]:
●The persistent wheeze phenotype was linked with parental history of asthma, non-White race/ethnicity, early-life bronchiolitis, and the genetic polymorphism ORMDL sphingolipid biosynthesis regulator 3 (ORMDL3), which is associated with increased risk of asthma.
●The mid-childhood wheeze trajectory was associated with atopic disease (eczema in infancy and elevated total serum IgE in adolescence).
●For the transient wheeze phenotype, bronchiolitis in infancy was the only associated risk factor.
ASTHMA RISK FACTORS — Genetic factors, perinatal exposures, and their interactions may contribute to the development of asthma. Other genetic, environmental, developmental, and host factors can alter the expression or persistence of the asthma phenotype over time. Some of these clinical indicators of risk are used in various predictive models to help the clinician identify those children who will continue wheezing into older childhood. Asthma risk factors are discussed in greater detail separately, but the key risk factors identified from the wheezing phenotypes studies are reviewed here. (See "Risk factors for asthma".)
Atopy — A variety of studies of wheezing phenotypes have evaluated the relationship between atopy and asthma [9,23-25]. The following observations illustrate the range of findings:
●Predisposition to allergy appeared to be the primary risk factor in the Tucson Children's Respiratory Study (TCRS) cohort described above for early-life wheezing leading to asthma [2]. Children in the late-onset and persistent wheezing phenotype groups had a significantly higher rate of allergen sensitization at six years of age than healthy children without wheezing (56 and 51 percent versus 34 percent, respectively). In addition, sensitization to the common mold Alternaria was associated with a higher rate of chronic asthma at age 22 years (odds ratio [OR] 3.6) [23]. (See 'Tucson classification' above.)
●The phenotypes in the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort described above that were most strongly associated with atopy were intermediate-onset, late-onset, and persistent wheeze [9]. These phenotypes also had the highest proportion of clinician-diagnosed asthma by 7.5 years of age. (See 'Avon classification' above.)
Differences in risk factors and outcomes are seen between nonatopic and atopic, persistent wheezing phenotypes:
●In children with the nonatopic, persistent wheezing phenotype, the initial wheezing episode occurs in the first year of life [26], and the wheezing episodes become less frequent by early adolescence [27]. Children with this phenotype have a lower level of prebronchodilator lung function and enhanced airway reactivity [27]. This phenotype is widespread among children with low socioeconomic status [28]. Other names, such as episodic (viral) wheeze and mild, early, viral wheeze, have also been used to describe this phenotype in children who principally wheeze with viral infection alone [17,29]. (See 'Symptom-based phenotypes' above.)
●In children with IgE-associated/atopic, persistent wheezing phenotype, wheezing generally starts after the first year of life and persists into later adolescence [2]. It remains to be demonstrated if children with this phenotype will continue to experience symptoms that persist into later adulthood. Risk factors linked with atopic wheeze include parental asthma, male sex, atopic dermatitis, eosinophilia at nine months, a history of wheezing with lower respiratory tract infections [2], early sensitization to food or aeroallergens [29,30], and the development of symptoms between wheezing exacerbations [17]. Children who are sensitized to more allergens are more likely to develop atopic, persistent wheezing [31].
The "atopic march" usually starts with atopic dermatitis in early life and progresses to the addition of other allergic diseases, including food allergy, allergic rhinoconjunctivitis, and asthma. Early age of onset of atopic dermatitis and allergic sensitization is associated with an increased risk of childhood asthma in several studies. (See "Atopic dermatitis (eczema): Pathogenesis, clinical manifestations, and diagnosis", section on 'Allergic rhinitis, asthma, and food allergy' and "Risk factors for asthma", section on 'Atopy'.)
Reduced lung function — Lung function is more impaired in children with persistent wheezing compared with other wheezing phenotypes [2,9,24,25,32,33]. In the Avon cohort, for example, the largest decrements in lung function were seen in the prolonged, early; intermediate-onset; and persistent wheezing groups, whereas airway hyperresponsiveness was greatest in the intermediate- and late-onset phenotypes [9].
The age of onset of reduced lung function has varied in different studies [2,23,24,32,34-36]. As examples:
●Lung function was normal in infancy but reduced at six years and into adolescence in children who developed persistent wheeze at six years in the Tucson group [2]. Both persistent wheezing in early life and airway hyperresponsiveness and low airway function at six years were associated with both chronic and newly diagnosed asthma at 22 years of age [23].
●Lung function was abnormal as early as one month of age in cohorts from Norway and Australia who developed persistent wheeze at ages 10 or 11 years [34,35].
Exposure to high levels of perennial allergens early in life has been associated with a greater decrease in lung function during the school years in children with atopic, persistent wheezing [37]. However, it is not known if these children had decreased lung function before the development of allergic sensitization.
Respiratory pathogens — Preschool wheezing illnesses with both viral and bacterial pathogens are also associated with an increased risk for recurrent wheeze and early childhood asthma. Colonization with common respiratory bacterial pathogens, such as Streptococcus pneumoniae, Moraxella catarrhalis, and Haemophilus influenzae [38,39], and viral respiratory infections in infancy, particularly with human rhinovirus [40,41], appear to act synergistically in increasing the asthma risk with both allergic sensitization and reduced lung function during early childhood [42,43]. Viral infections can lead to changes in the composition of the airway microbiome and overgrowth of respiratory pathogens, likely contributing to respiratory symptoms and perhaps airway obstruction [39,44]. Although bacterial pathogens are linked to wheezing and asthma, environmental exposure to a broad range of bacteria might protect against wheezing illnesses [45,46]. (See "Role of viruses in wheezing and asthma: An overview", section on 'Development of asthma' and "Risk factors for asthma", section on 'Respiratory infections'.)
Microbiome imbalance — An imbalance in the gut microbiome is associated with both the development of atopic diseases and a heightened predisposition to viral infections [47]. It is postulated that colonization with protective gut bacteria has a direct anti-inflammatory effect on the respiratory tract, reducing airway hyperreactivity [48]. Other species of gastrointestinal microbiota are associated with an increased ability of the respiratory immune system to combat viral pathogens [48]. It is also hypothesized that exposure to diverse microbes and allergens prenatally or during infancy is immunostimulatory and inversely related to the risk of wheezing illnesses [45,49,50]. Moreover, reduction of the immune and inflammatory processes during respiratory infections is another potential pathway for asthma prevention. As an example, a randomized trial of 40 infants hospitalized with respiratory syncytial virus (RSV) bronchiolitis treated with two weeks of azithromycin during the acute illness demonstrated less frequent wheeze during the subsequent year in treated children (36.7 days compared with 70.1 in the placebo group) [51]. The investigators attributed the effects of azithromycin to antiinflammatory properties or modification of the airway microbiome [52].
Vitamin D — Vitamin D is postulated to promote in utero lung growth and the development and enhancement of antimicrobial effects, leading to reduced early-life respiratory infections and/or providing immune modulation effects [53]. A meta-analysis of two trials supports the concept that maternal vitamin D intake is protective for the occurrence of asthma in offspring during the first three years of life, particularly in females with normal serum vitamin D levels (≥30 ng/mL) at randomization [54].
Recurrent asthma symptoms — A history of clinician-diagnosed recurrent wheezing in preschool-aged children is associated with an increased risk of asthma [2,55]. Results from one prospective study suggest that quantitative measures of family-reported recurrent asthma symptoms (eg, noisy breathing, shortness of breath, or persistent troublesome cough) may be better predictors of asthma than clinician-diagnosed wheezing episodes alone [56].
PREDICTIVE TOOLS IN CHILDREN WITH WHEEZING — Various predictive models or clinical indicators of risk have been studied to help the clinician identify those children who will continue wheezing into older childhood. These models have employed various risk factors associated with the development of asthma in epidemiologic studies, such as parental history of allergic sensitization and asthma, wheezing history, atopic disease in the child, IgE levels, and cytokine secretion profiles. However, none of these clinical tools have been validated in populations different from the study group.
A 2015 systematic review of 12 childhood asthma prediction models that assessed symptomatic children up to four years of age and predicted subsequent development of school-age asthma revealed the following predictive factors: demographic factors, respiratory symptoms, number of respiratory tract infections or wheezing episodes and hospitalizations, family history of allergy or asthma, other comorbid allergic conditions, eosinophilia, total IgE, specific IgE and allergen skin-prick testing to both food and aeroallergens, fraction of exhaled nitric oxide (FeNO) levels, and preterm or post-term delivery [57]. No model reviewed had both high sensitivity and specificity (good at ruling out and ruling in disease).
Predictive scoring systems — Several asthma risk scores have been developed for use in children with at least one episode of wheezing in early childhood. Most of the risk factors included in these scores are easily determined from the patient history and physical exam.
Asthma Predictive Index (API) — The API was derived from an unselected multiethnic population of children in the Tucson Children's Respiratory Study (TCRS) cohort who had wheezed at least once during the first three years of life [58-60]. The major criteria were clinician-diagnosed eczema or parental asthma. The minor criteria were clinician-diagnosed allergic rhinitis, wheezing apart from colds, and eosinophilia ≥4 percent. A positive loose index was defined as any parental/caregiver report of wheezing on the surveys at two or three years of age and either one major criterion or two minor criteria. A positive stringent index was defined as frequent wheezing on these same surveys (score of ≥3, scale: 1 to 5, from "very rarely" to "on most days") plus the same combination of major or minor criteria. Children with a positive loose index were four times more likely to have active asthma during a subsequent survey at 6, 8, 11, or 13 years of age (sensitivity 42 percent, specificity 85 percent). Children with a positive stringent index were seven times more likely to have active asthma in at least one of these school-aged surveys (sensitivity 16 percent, specificity 97 percent). In addition, the positive likelihood ration (LR), the probability of a child with active asthma classified as at risk divided by the probability of a child without active asthma classified as at risk, of the API is 7.43 at age six years [61]. The negative LR, the probability of a child with active asthma classified as not at risk divided by the probability of a child without active asthma classified as not at risk, is 0.75 at age six years [61].
The API was validated in a population-based birth cohort in Leicester, United Kingdom [62]. The API and wheeze frequency were assessed in 1954 children at three years of age and then compared with rates of asthma at 7 and 10 years of age. Results were similar to the findings in the Tucson cohort, with a fivefold increased risk of asthma at seven years of age with a positive loose API and an eightfold increased risk with a positive stringent API. Using a simpler rule of early wheeze (wheezing in the first three years of life) or early, frequent wheeze yielded comparable results. Thus, using the simpler definition may save screening costs. However, the overall predictive performance for all measures was low [63,64].
Yet, there remains a need to identify children that are unlikely to develop later asthma in order to prevent exposing children inappropriately to costly preventive therapies with potential side effects. The sensitivity of the API index is low (15 to 57 percent), suggesting that the test is poor at predicting later asthma [58-60,62,65,66]. However, the API has a high negative predictive value in the populations tested, which is essential for identifying children that have a low probability of having later asthma when the API is negative [23,59,60]. (See "Glossary of common biostatistical and epidemiological terms" and "Evaluating diagnostic tests".)
A modified version of the API, which increases the number of wheezing episodes from three or more to four or more and replaces provider diagnosis of allergic rhinitis with allergic skin testing, has been endorsed by the US National Asthma Education and Prevention Program Expert Panel Report 3 for use in the diagnosis of asthma [67]. It may also be useful for identifying treatment responders, as demonstrated in the Prevention of Early Asthma in Kids (PEAK) study [68]. The predictive value of this modified API was tested in a cohort of high-risk children with a family history of allergy and/or asthma [69]. A positive modified API (mAPI) increased future asthma probability from a pretest probability of 30 percent to a posttest probability of 90 percent. The positive LR of the mAPI in this high-risk population is 21 and the negative LR is 0.84 at six years of age [61].
Another modified version of the API using two or more wheezing episodes and keeping the other criteria similar to that used in the PEAK study was tested in a cohort of high-risk children with a family history of allergic sensitization [70]. Asthma at age seven years was significantly associated with this modified API (adjust odds ratio [OR] 13.3, 95% CI 7.0-25.2). The sensitivity of this score was 44 percent, specificity 94 percent, positive LR 7.5, and negative LR 0.6 at seven years of age [61].
The Asthma Detection and Monitoring (ADEM) API [71] was based upon a prospective case-control, population-based study of 219 infants with history of recurrent wheezing. The original API was modified by replacing eosinophilia with specific IgE and the addition of 17 exhaled volatile organic compounds, 10 exhaled breath condensate biomarkers (cytokines and chemokines), genes expression, and lung function. The sensitivity of this score was 88 percent, specificity 90 percent, positive LR 8.8, and negative LR 0.13 at six years of age [61]. This model is the most complete and sophisticated model and reaches the best positive and negative LRs; however, it may be costly to implement in general practice.
Other asthma risk scores
Isle of Wight score — Information collected prospectively from an Isle of Wight whole population birth cohort identified four factors in a logistic regression model that were independently predictive of asthma [72]. These factors included recurrent chest infections at two years of age and family history of asthma based on questionnaire data, skin prick test positivity to at least one food and/or inhalant allergen at four years of age, and recurrent nasal symptoms at one year. There was a nearly ninefold increased risk of persistent wheeze at 10 years of age in children positive for all four risk factors with a history of at least one episode of wheezing in the first four years of life. Wheezing persistence at 10 years of age was identified in 83 percent of children with all four risk factors, and wheezing transience was identified in 80 percent with none of these risk factors. The sensitivity of this score was 72 percent, specificity 71 percent, positive LR 2.5, and negative LR 0.4 at 10 years of age [61].
Leicestershire tool — Another asthma prediction tool was developed using a population-based cohort of 1226 children from Leicestershire, England, aged one to three years, with history of cough or wheeze [73]. The diagnosis of asthma five years later was predicted using a tool consisting of 10 variables representing wheeze severity and triggers, male sex, age, eczema, and parental history of asthma (total score range, 0 to 15). The probability of asthma was 95 percent for a score of 15, 72 percent for 10, and 28 percent for 5. The sensitivity of this score was 53 percent, specificity 85 percent, positive LR 3.41, and negative LR 0.56 at 10 years of age [61].
Persistent Asthma Predictive Score (PAPS) — This tool was developed using a retrospective analysis of clinical and biologic data of 200 children less than two years of age with recurrent wheezing and evaluation of current asthma in these same children at six years of age using the International Study of Asthma and Allergies in Childhood (ISAAC) questionnaire [74]. Three variables (family history of asthma, personal history of atopic dermatitis, and sensitization to multiple allergens) were included in the score, which had a positive predictive value of 67 percent and a negative predictive value of 76 percent. This tool was validated in a second cohort of 227 children.
Environment and childhood asthma score — This score was based upon the severity of obstructive airway disease (number of episodes, months of persisting symptoms, and number of hospitalizations for bronchial obstruction) during the first two years of life in a nested case-control study within a Norwegian birth cohort [75]. Children with bronchial obstruction during the first two years were at significantly higher risk of asthma at 10 years of age than those with no obstruction (36 versus 6 percent, respectively). There was a linear association between the severity score (range, 0 to 12) and risk of asthma at 10 years. Children with a score of 6 to 12 were 20 times more likely to have asthma at 10 years of age than those without a history of bronchial obstruction in the first two years of life.
PIAMA risk score — Risk factors associated with asthma at seven to eight years were identified in a subset of higher-risk children with at least one episode of wheezing or nocturnal cough without a cold in the first four years of life in the Prevalence and Incidence of Asthma and Mite Allergy (PIAMA) birth cohort in the Netherlands [76]. A clinical risk score (range, 0 to 55) was developed using eight weighted risk factors: male sex, postterm delivery, medium/low parental education, wheezing frequency (one to three or four or more times per year), wheezing/dyspnea apart from colds, parental report of serious infections (one to two or three or more times per year), and presence of clinician-diagnosed eczema. Children with a score of 30 or greater had an increased risk of asthma at seven to eight years of age compared with those with a score of 10 or less (42 versus 3 percent, respectively). The PIAMA risk score was validated in a study of 2877 children participating in a multiethnic, population-based cohort who had asthma-like symptoms at preschool age and were reassessed for asthma at six years of age [77]. The sensitivity of this score was 60 percent, specificity 76 percent, positive LR 2.5, and negative LR 0.53 at seven to eight years of age [61].
Children from the PIAMA study discussed previously were recruited for further testing at four years of age if they had a positive response to at least one respiratory symptom suggestive of asthma (wheeze, shortness of breath, or nocturnal cough without a cold) in the past 12 months on the three- and four-year questionnaires [78]. Higher FeNO and sensitization to at least one of six inhalant allergens (specific IgE) were positively associated with wheezing and asthma at eight years of age (OR 1.6 for FeNO and 2.8 for specific IgE), independent of clinical history at four years of age. The combination of FeNO, specific IgE, and clinical history (wheeze one to three or four or more times, eczema, and atopic mother) improved the prediction of wheezing at eight years of age compared with each of these measures alone.
Clinical Asthma Prediction Score (CAPS) — Asthma symptoms and environmental conditions were assessed and specific IgE testing to cat, dog, and house dust mite was performed at baseline in 771 Dutch preschool children with a history of recurrent cough, wheezing, or shortness of breath seen in a primary care clinic [79]. Patients were reassessed at six years of age. Five variables predicted asthma at six years of age: older age at presentation, family history of asthma or allergy, wheezing-induced sleep disturbances, wheezing apart from colds, and specific aeroallergen IgE >0.35 kU/L. The scores ranged from 0 to 11 points; scores <3 were associated with a negative predictive value of 78 percent, and scores >7 were associated with a positive predictive value of 74 percent.
Pediatric Asthma Risk Score (PARS) — Demographic and clinical data from 762 children from a birth cohort enrolled in the Cincinnati Childhood Allergy and Air Pollution Study were used to identify factors that predicted asthma development [80]. PARS predicted asthma development in the Cincinnati cohort (sensitivity 0.68, specificity 0.77) and was replicated in the Isle of Wight birth cohort (sensitivity 0.67, specificity 0.79). Variables that predicted asthma in PARS included parental asthma, eczema, wheezing apart from colds, early wheezing, sensitization to two or more food allergens and/or aeroallergens, and African American race.
FUTURE DIRECTIONS — Multidimensional approaches that include novel variables, such as biomarkers [81-87], lung imaging [88,89], genetics [22,90-95], and statistical techniques [96], may improve on the fidelity of phenotype designations and their stability over time. With further study, phenotypes may be identified that are highly predictive of the ultimate expression of asthma. This in turn may improve the precision of research into mechanisms causing asthma, with the ultimate goal of the primary prevention of asthma.
SUMMARY
●Several early-childhood wheezing phenotypes have been described based upon the natural history and associated risk factors. Accurate classification of young children at high risk to develop persistent asthma may help predict long-term outcomes and identify children who may benefit from secondary prevention interventions. (See 'Introduction' above.)
●Wheezing phenotypes have been defined to identify the characteristics and risk factors associated with children that experience wheezing. These include several epidemiologic phenotypes that are based upon wheezing history and symptom-based phenotypes that are based upon temporal patterns. The most well-known of these phenotypes is the classification from the Tucson Children's Respiratory Study (TCRS) cohort, which divided young children with wheezing into the following groups: early, transient wheezers; persistent wheezers; and late-onset wheezers. The European Respiratory Society defined two symptom-based phenotypes: episodic (viral) wheeze and multi-trigger wheeze. (See 'Wheezing phenotypes' above.)
●Genetic, environmental, developmental, and host factors may contribute to the development of asthma and can alter the expression or persistence of the asthma phenotype over time. The major risk factors identified in various wheezing phenotype studies include the presence of atopy (in both the child and the parents), reduced lung function, and viral infections in infancy, particularly with rhinovirus. (See 'Asthma risk factors' above and "Risk factors for asthma" and "Role of viruses in wheezing and asthma: An overview".)
●Various predictive models or clinical indicators of risk have been studied to help the clinician identify those children who will continue wheezing into older childhood. These models have employed various risk factors associated with the development of asthma in epidemiologic studies, such as parental history of allergic sensitization and asthma, wheezing history, atopic disease in the child, immunoglobulin E (IgE) levels, and cytokine secretion profiles. The most commonly used predictive scoring system is the Asthma Predictive Index (API). (See 'Predictive tools in children with wheezing' above.)
