Signs and Symptoms that Precede Wheezing in Children with a Pattern of Moderate-to-Severe Intermittent Wheezing

Methods 

Children age 12 to 59 months (n = 238) were randomized in the Acute Intervention Management Strategies (AIMS) trial, a multicenter, double-blind, placebo-controlled study performed by the National Health Lung and Blood Institute–sponsored Childhood Asthma Research and Education (CARE) Network. The inclusion criteria for the trial were aimed at identifying children who experienced 2 or more episodes of moderate to severe wheezing in the preceding year in the context of an upper respiratory tract illness but who remained well between episodes and did not demonstrate features of “persistent asthma.” Participants randomly received inhaled corticosteroid (budesonide), leukotriene receptor antagonist (montelukast), or placebo for 7 days at the onset of each RTI for a period of 1 year. All participants received the study medication plus albuterol inhalation treatments 4 times daily while awake (and more as needed) for the first 48 hours after the onset of an RTI, followed by albuterol inhalation on an as-needed basis. Parents were instructed to begin study medications as soon as the child-specific symptoms were detected. Each center's Institutional Review Board approved the study design, and all parents signed informed consent.1

Parents completed the initial surveys at the time of randomization under the supervision of a clinical coordinator. They subsequently completed identical surveys at home at the onset of each subsequent RTI during the trial. They reviewed their responses with the clinical coordinators on the phone within 48 to 72 hours of initiation of study medication, to discuss the scenario that prompted this initiation, as well as the course of the illness. These surveys identified the signs and symptoms that preceded the eventual development of wheezing, detailing various symptom categories as well as specific symptoms within each category (Table I). Parents identified the very first symptom that they noticed as part of a RTI likely to include wheezing, the most important symptom that made them certain of the subsequent development of extensive wheezing, and the symptom that led them to initiate treatment (study medication). The survey was developed and refined based on the results of a pilot study of 28 parents of children with the same inclusion criteria described earlier that was completed before the clinical trial at each clinical center. Parents also were given a diary card on which to record respiratory symptoms on a twice-daily basis, to supplement the data that they provided on the illness survey.

Table I. RTI survey and symptoms included in the survey

	A. Questions
	1. What is usually the very first symptom you notice that leads you to believe your child is starting a breathing illness?
	2. What is usually the most important symptom you notice that makes you feel very certain that the illness will lead to significant breathing problems?
	3. What 2 symptoms are usually present when you first give medications intended to lessen the symptoms of your child's breathing illness?
	B. Symptoms included in the survey

	General symptom category	Specific signs and symptoms
	Appearance changes	Dark circles under eyes, glassy eyes, watery eyes
	Appetite changes	Eating less/will not eat, spitting up/vomiting
	Behavior problems	Bedwetting, fussy/cranky/irritable, hyperactive, less active (will not play)
	Breathing problems	Breathing worse, cannot breathe, nasal flaring, not breathing well, pulling of ribs/neck, rapid breathing, shortness of breath, color change, turning blue
	Changes in sleep patterns	Awakening, sleepy during the day/lethargic
	Insignificant cough	Infrequent, mild, not concerning
	Significant cough	Concerning, constant, interrupts activity, interrupts sleep, repetitive, “asthma cough”
	Fever	Any fever, high fever, skin feels warm/hot to touch
	Noisy breathing	Hoarse voice, snoring
	Noisy chest	Gurgling, rattling, wheezing

Results 

Parent and Patient Characteristics 

A total of 238 children were randomized between March and November 2004. Of these, 215 children experienced at least 1 RTI and 176 experienced 2 or more RTIs during the 12-month trial. All parents completed the survey at the time of randomization and then were instructed to complete the survey at the onset of each subsequent RTI. Parents completed surveys for 194 first RTIs (90%) and 148 second RTIs (84%).

More than 50% of participating parents had at least some college education, and the majority had a household income of $50 000 or more (Table II; available at www.jpeds.com). The survey was completed predominantly by mothers (96%). The majority of the children were male (64.7%) and Caucasian (55.9%), and more than 70% experienced at least 4 wheezing episodes over the preceding 12 months and did not experience active symptoms during the month preceding enrollment or during a 2-week prospective run-in period.1 Additional characteristics of the children randomized into the trial have been published previously.1

Table II. Parent demographics and child characteristics

	A. Parent demographics
	Sex
	Female	95.8%
	Male	4.2%
	Age, years
	19 or younger	1.0%
	20-29	32.1%
	30-39	55.4%
	40-49	10.1%
	Highest education
	Did not complete 8th grade	0.8%
	Completed 8th grade	1.3%
	Some high school	8.8%
	Completed high school	11.8%
	Some college/postsecondary education	29.4%
	Completed college	28.6%
	Completed higher education	19.3%
	Annual household income
	Less than $15 000	13.3%
	$15 000-$29 999	12.4%
	$30 000-$49 999	18.8%
	$50 000 or more	55.4%
	B. Characteristics of children enrolled in the trial (AIMS cohort)
	n	238
	Age, years	3.0±1.1
	Sex, % male	64.7
	Race, % Caucasian	55.9
	Courses of oral corticosteroids in past year, n	1.2±1.4
	Emergency department/hospital visits in past year, n	1.1±2.2
	Doctor visits in past year, n	4.1±3.1
	Days missed from school/daycare, n	4.8±8.9

Data are expressed as mean ± SD except when noted otherwise.

Symptoms 

“Nose symptoms” and “significant cough” (cough that is concerning, constant, interrupts activity, interrupts sleep, repetitive, or described as the “asthma cough”) were the first symptoms of a RTI reported most often at randomization and at the onset of the first 2 RTIs (Figure, A). As participants experienced increasing numbers of RTIs during the trial, nose symptoms remained the most prevalent very first symptom, but the frequency with which significant cough was reported as the very first symptom increased significantly (Figure, A; P = .03 for trend). No significant change in the frequencies of the 4 other major symptom categories (“insignificant cough,” “noisy chest,” “breathing problems,” and “behavior changes”) was seen over the first 2 RTIs.

View Large Image Download to PowerPoint

Figure. Histograms representing the frequencies of symptoms reported by parents as the very first symptom A, the most important symptom B, and the symptom initiating treatment of an RTI C.

The most commonly reported response in the most important symptom category at randomization and during the first and second RTIs was significant cough (in 40%, 50%, and 53% of participants, respectively) (Figure, B). Although the frequency of “significant cough” as the most important symptom increased numerically between randomization and the second RTI, this change was not statistically significant. Similarly, there was no significant change in the frequencies of the 5 other major symptom categories for the most important symptom.

Significant cough was the most common symptom for which parents initiated study medication at randomization, first RTI, and second RTI (in 42%, 46%, and 51% of participants, respectively) (Figure, C). There was a numerical increase in the proportion of parents who initiated treatment for nose symptoms as the number of RTIs increased, along with numerical decreases in initiation of study medication due to noisy chest or breathing problems; trend analysis found that these differences were not statistically significant, however.

The 3 study groups exhibited no differences in symptom reporting patterns treatment across RTIs for the very first symptom (P = .65), most important symptom (P = .57), or symptom initiating treatment (P = .53). Early symptom patterns did not influence episode severity, as reflected by oral corticosteroid use; the early symptom patterns during RTI episodes that led to oral corticosteroid use did not differ from those during RTIs that did not require oral corticosteroid use (P = .88).

As reported on the diary cards, 69% of participants developed wheezing during the first RTI. Table III summarizes the performance characteristics of individual symptoms reported as the very first, most important, and reason for initiating treatment in the first RTI. For the very first symptom, parental identification of significant cough was highly specific (78%) and was associated with a PPV of 74% in predicting wheezing during the RTI. When reported as the most important symptom, noisy chest was associated with very high specificity (97%) and PPV (90%) for the development of wheezing; however, this finding is confounded by wheezing as a specific symptom within the noisy chest category, and although noisy chest is very specific, it has low sensitivity. Significant cough had the highest sensitivity (47%) and PPV (62%) when identified as the most important symptom. When identified as the symptom that initiated treatment, significant cough had a sensitivity of 48% and a PPV of 68% for predicting subsequent wheezing, in contrast to insignificant cough, which had a very low sensitivity (8%) but a high PPV (71%). Treatment group did not influence the performance characteristics described earlier, because there were no significant differences in sensitivity or specificity by treatment group (all P > .05).

Table III. Performance characteristics of symptom categories in predicting wheezing during first RTI

		Symptom	Sensitivity	Specificity	PPV	NPV
	Very first symptom	Significant cough	33%	78%	74%	37%
		Insignificant cough	13%	88%	67%	34%
		Nose	37%	50%	59%	29%
		Noisy chest	3%	100%	100%	35%
	Most important symptom	Significant cough	47%	45%	62%	31%
		Nose	11%	81%	54%	32%
		Noisy chest	15%	97%	90%	37%
	Symptom initiating treatment	Significant cough	48%	56%	68%	36%
		Insignificant cough	8%	94%	71%	34%
		Nose	24%	66%	57%	31%
		Noisy chest	9%	100%	100%	36%

NPV, Negative predictive value.

Consistency of Responses Between RTIs 

Significant cough was the symptom most likely to be identified consistently between RTIs. When significant cough was reported as the most important symptom during the first RTI, 62% of parents reported this category as the most important symptom during the second RTI (Table IV). Other symptom categories had lower concordance rates between the first 2 RTIs. This pattern was similar to that observed between the symptom category identified based on previous history (at randomization) and the symptom category reported during the first RTI (Table V; available at www.jpeds.com).

Table IV. Concordance between survey responses during respiratory tract illness (RTI)

	Symptom		Rate of concordance first RTI to second RTI
	1st RTI	2nd RTI	Very first symptom	Most important symptom	Symptom to treatment
	Nose	Nose	47%	9%	42%
	Significant cough	Significant cough	43%	62%	64%
	Insignificant cough	Insignificant cough	—	8%	9%
	Insignificant cough	significant cough	38%	25%	36%
	significant cough	Insignificant cough	10%	8%	4%
	Noisy chest	Noisy chest	—	23%	—
	Behavior Change	Behavior Change	33%	20%	—

Table V. Concordance between survey responses during retrospective versus prospective RTI

	Symptom		Rate of concordance randomization to first RTI (prospective to retrospective)
	Randomization	First RTI	Very first symptom	Most important symptom	Symptom initiating treatment
	Nose	Nose	63%	17%	44%
	Significant cough	Significant cough	49%	59%	54%
	Insignificant cough	Insignificant cough	22%	23%	12%
	Insignificant cough	Significant cough	39%	39%	36%
	Significant cough	Insignificant cough	10%	10%	5%
	Noisy chest	Noisy chest	10%	18%	9%
	Behavior change	Behavior change	39%	36%	5%

Discussion 

Recognizing the variability in patterns of symptom development among patients is imperative for developing appropriate management strategies for children with moderate to severe intermittent wheezing. Thus, clinicians must understand these symptom patterns and help parents of these children recognize such patterns, to ensure timely and effective treatment. But this is complicated by imperfect parental identification of wheezing, because clinicians depend heavily on reported parental impressions and perceptions.4, 5 In preschool-age children with a history of moderate to severe intermittent wheezing enrolled in the AIMS trial, nose symptoms were the most commonly recognized first symptom at the onset of a RTI that parents suspected would eventually include wheezing; however, nose symptoms had low sensitivity and PPV for subsequent wheezing. Although reported less frequently than nose symptoms as the very first symptom of RTI, significant cough was most often identified as the most important symptom predictive of wheezing, although it had relatively low sensitivity, but fair PPV, for predicting wheezing.

These findings suggest that parents' perceptions of the development of particular respiratory symptoms at the onset of a RTI, particularly significant cough, have acceptable PPVs and specificities for predicting subsequent wheezing after the onset of a RTI, and may serve as indicators for escalation of therapy in an attempt to minimize subsequent wheezing. In other words, waiting until wheezing occurs to initiate therapy is too late. The clinician should discuss the specific characteristics of the cough with their patient's parents during asthma management planning. These findings also suggest that even though significant cough was the symptom category most predictive of subsequent wheezing, each child needs to be treated individually, because not all parents identified significant cough as the most important symptom or the symptom initiating treatment.

Our findings may be of critical importance when conducting clinical trials involving management of RTIs in children with the aim of preventing wheezing and associated morbidities. Over the past 2 decades, multiple clinical trials in children have targeted this aim.11, 12, 13, 14, 15, 16, 17, 18, 19, 20 These trials used various approaches to trigger initiation of study intervention, including “first sign of an attack,”11 “onset of upper respiratory tract symptoms which typically precipitated asthma attacks,”12 “first sign of upper respiratory tract infection,”16 and “increased in symptom scores by one point in a symptom scale.”19 It may be that a lack of attention to the clinical features specified in our study led to submaximal responses in those trials. We propose that improving our understanding of the early changes that occur before a wheezing illness in each individual patient can allow us to better educate the parents about when to initiate a specific intervention.

The present study has some limitations to consider. First, because parents completed the survey during the course of a clinical trial under close monitoring, their responses may be the result of increased vigilance and might not be generalizable to the usual clinical setting, in which supervision is less intense. Most of the parents identified significant symptoms instead of trivial or controversial signs of RTI. It is our impression that most parents of children with moderate to severe recurrent wheezing are very attentive to their children during RTIs and thus can be expected to provide accurate responses to similar questions. Moreover, the parents were well educated; how these results would translate in a less-educated group is not clear. There could be a parental learning effect as the number of RTIs progresses, such that ongoing guided observation may lead to a shift between symptom categories. Nevertheless, as described earlier, the consistency between responses for significant cough between RTIs was 43% for the first symptom and 62% for the most important symptom. This suggests that even though some parents recognized different symptom patterns with subsequent RTIs, most of them recorded similar symptom observations across RTIs when reporting significant cough. Our definition of noisy chest included wheezing, which likely accounts for the very high specificity and PPV. In addition, it is likely that study treatment(s) had an effect on the appearance of wheezing; our study was not designed to investigate this question, however. Our analysis is based on parental reporting and perception of symptoms with the aid of a coordinator and/or physician assistance to decide RTI management. But the children did not undergo a physician's examination with each RTI, and thus wheezing was not objectively documented. However, it is common clinical practice to provide an “action plan” for parents to initiate home-based treatment for wheezing without physician examination.

In summary, the presence of significant cough as an early sign of RTI is a frequently reported and reliable predictor of subsequent lower RTI associated wheezing in preschool children with a history of severe intermittent wheezing. But here “one size does not fit all;” several other symptom categories were frequently identified as antecedents of wheezing episodes. This suggests the need for further investigation of individualized management strategies that will be most effective for each child.