INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bronchial hyperresponsiveness is considered one of the key features in pediatric asthma (1). Bronchoconstriction induced by hyperventilation of cold, dry air is commonly found in asthmatic adults and children (2). The mechanism of action is thought to be respiratory heat and water loss with transient hyperosmolarity in the respiratory mucosa causing the airways to narrow. Whether the magnitude of airway obstruction is determined by airway cooling alone or by evaporation of water from the respiratory mucosa irrespective of cooling, is a matter of debate (9, 10). Cold, dry air challenge (CACh) has two main attractive features: it is simple to perform and standardize even in young children 2 to 5 yr of age, and the reaction probably reflects the pathophysiology of asthma better than a pharmacological bronchoprovocation with histamine or methacholine (11). Cold air challenge in young children of 2 to 5 yr has not previously been reported, since no acceptable method of lung function testing has been available in awake young children.

We have recently evaluated lung function techniques that allow measurements during tidal breathing in nonsedated young children as young as 2 yr of age (12). The techniques comprise specific airway resistance (sRaw) by whole body plethysmography; respiratory resistance by the interrupter technique (Rint); respiratory reactance at 5 Hz (Xrs5); and respiratory resistance at 5 Hz (Rrs5) by the impulse oscillation technique.

The aims of the present study were: (1) to study feasibility of CACh in young children 2 to 5 yr of age, and (2) to determine sensitivity, specificity, predictive values, and repeatability of CACh in order to evaluate if CACh could distinguish asthmatics from healthy children in this age group using sRaw as primary outcome.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Asthmatics. Young children age 2 to 5 yr with asthma were eligible for the study. They were all recruited from our pediatric outpatient clinic. All had recurrent asthmatic symptoms and were being treated for asthma. Regular treatment was stopped and terbutaline as required was the only allowed medication during an observational period of 2 wk to a maximum of 8 wk in which asthma symptoms were registered in a diary. Patients with daily symptoms in at least 7 of 14 consecutive days were scheduled for a CACh test. In half of the patients another CACh was scheduled after 8 wk during which terbutaline as required was the only allowed medication. On the day of CACh they had no asthma symptoms. All asthma medications were stopped 12 h before CACh.

Controls. Children without any previous lung illness, atopic symptoms, atopy in first-degree relatives, or smoking at home were considered eligible for recruitment into the study. The control group was collected by a health questionnaire which was mailed to families of young children 2 to 5 yr of age living in Greater Copenhagen. The families were randomly selected from the municipal population database. Table 1 lists baseline characteristics of the study groups. The parents of all children gave written informed consent to the study protocol, which was approved by the local ethics committee of Copenhagen (KF 02-115/96).

Pulmonary Function Testing

Pulmonary function testing was performed with a Master Screen unit, version 4.22 (E. Jaeger GmbH, Würzburg, Germany). Methods and equipment have previously been described in detail (12, 13). Measurements were done during tidal breathing in a face mask (Astratech No. 2; ASTRA, Albertslund, Denmark) fitted with a flexible, noncompressible mouthpiece securing support of cheeks, mouth breathing, and stable access to the airways (12). Flow and volume were measured with a heated pressure screen-type pneumotachograph with a resistance of 0.036 kPa · s · L¹. The equipment was calibrated daily.

sRaw. sRaw was measured as the relationship between simultaneous variations of respiratory flow and variations of pressure in a constant-volume whole body plethysmograph. This relationship was presented on the on-line monitor as s-shaped loops. sRaw was calculated at the maximal changes in plethysmographic pressure during inspiration and expiration (14). BTPS compensation (body temperature, barometric pressure, and saturated with water vapor) was electronic. The child was coached to reach a respiratory rate of 30 to 45 breaths/ min. When needed, the child was accompanied by an adult (15). Any disturbance in measurements from face mask leakage, coughing, swallowing, or vocalization was seen as an alteration in the shape of the resistance loops and resulted in repeated measurements. The median value of five sequential specific resistance loops, similar appearing as judged from slope and shape, was retained as outcome.

Rint. Respiratory resistance measured by the interrupter technique is based on the assumption of a simple relation between mouth pressure at the end of interruption of airflow and the airflow after reopening of a shutter valve mounted on the pneumotachograph. At every second inspiratory phase, inspiration of 50 ml activated the shutter for 80 ms. Mouth pressure was measured during the last 5 ms of the interruption. Flow was measured over 70 ms after reopening. The mean value of five consecutive, technically satisfactory measurements (identical pressure and flow curves) was saved as result. Deviation of values more than ± 15% from the mean value, dampening of pressure oscillation as seen during face mask leakage and swallowing, extreme pressure oscillations as seen during coughing, or ragged oscillations as seen during vocalization resulted in a new series of measurements.

Xrs5 and Rrs5. Xrs5 and Rrs5 were measured by the impulse oscillation technique (IOS). Rectangular impulses were generated mechanically by a loudspeaker and applied to the respiratory system through the mouthpiece. The resulting pressure and volume signals are analyzed for amplitude and phase difference to determine Xrs and Rrs of the respiratory system. Thirty seconds of undisturbed breathing at a respiratory frequency within 20 to 40 breaths/min resulting in technically satisfactory measurements were used as outcome (12, 13). Disturbance from coughing or vocalization resulted in either a new measurement or prolongation of data acquisition until 30 s of undisturbed measurement was obtained.

Sequence and timing of measurements. In a pilot study in 14 asthmatic schoolchildren performing spirometry 3, 4, 6, and 8 min after CACh we found maximal response in the period 3 to 5 min postchallenge. Therefore response to challenge was measured in the time interval between 3 and 5 min after the end of a challenge. As baseline values, duplicate measurements by each method were used before CACh. The sequence was: Rint, Xrs5, Rrs5, and sRaw. Measurements after CACh started with duplicate measurement of sRaw followed by single determinations of Xrs5, Rrs5, and Rint.

CACh

Cold, dry air was generated by a Respiratory Heat Exchange System (RHES; E. Jaeger GmbH, Würzburg, Germany). A schematic presentation of the equipment used in the study is shown in Figure 1.

View larger version (31K): [in this window] [in a new window]   Figure 1.   Schematic illustration of the cold air challenge equipment. Via face mask and mouthpiece the child hyperventilated cold, dry air mixed with 5% CO2 from the reservoir balloon. Mixing of dehumidified air and 100% CO2 was adjusted by flow meters (FLOW). Mixed air was cooled via the heat exchanger. Hyperventilation was measured via the pneumotachograph (PNT) and this analog signal was converted into a digital signal by the analog-digital converter (AD-CONVERTER) and displayed as a balloon on the computer screen. The child was told to keep the balloon flying just above the preset dashed line until the sun had crossed the screen. T = thermometer.

The system was set to deliver 15° C cold, dry air mixed with 5% CO₂. Temperature was measured 10 cm from the mouth. The test was carried out as a single-step 4-min isocapnic hyperventilation test. The child was breathing through the face mask fitted with a mouthpiece. This effectively secured mouth breathing and prevented inhalation of room air. The level of hyperventilation was aimed at 1 L/min/kg body weight. The flow of dry air and 100% CO₂ was mixed from separate flow meters to secure that total airflow was in accordance with the aimed hyperventilation rate and contained 95% dry air and 5% CO₂. A balloon served as reservoir for the mixed air. A pilot study was undertaken in which transcutaneously measured CO₂ in blood (Ptc_CO2) was monitored during CACh test. Ptc_CO2 was measured by a blood-gas monitor, Radiometer TCM 3 (Radiometer, Copenhagen, Denmark). Calibration was performed with a calibrating gas containing 20.9% O₂ and 5% CO₂. The temperature of the electrode was 44° C during measurements. All measurements were corrected to 37° C. The electrode was placed on the flexor side of the left forearm. The electrode was allowed to stabilize until steady-state readings for at least 5 min. The mean values of Ptc_CO2 were calculated from four readings at 1-min intervals of the hard-copy curves printed during measurements, before and during CACh, respectively. Eight children were investigated. None of these children participated in the actual study. Mean age of these children was 4.5 yr, and mean weight 18.3 kg. Mean baseline Ptc_CO2 was 4.8 kPa and mean Ptc_CO2 during CACh test was 5.1 kPa, i.e., a mean (range) increase in Ptc_CO2 of no more than 0.3 (0-0.5) kPa. Ptc_CO2 did not decrease in any of the children. Our method of hyperventilation of air mixed with 5% CO₂ in this age group was therefore considered safe.

Ventilation was measured at the exhalation valve (Hans Rudolph, Inc., Kansas City, MO) by a pneumotachograph (Hans Rudolph, Inc.). The analog signal from the pneumotachograph was converted by an analog-digital converter (Pico Technology Limited, Cambridge, UK) and displayed as the flying height of a computer-animated balloon (Figure 1). A horizontal dashed line on the computer screen was adjusted to reflect the target ventilation rate. The child was coached to maintain hyperventilation at a rate keeping the balloon above the predetermined level for 4 min. The child was able to keep track of time from a computer-animated sun, which crossed the screen in 4 min (Figure 1). All measurements at baseline and after CACh were performed by one observer.

Data Analysis

Response to CACh was quantified by the change from baseline values calculated as numbers of within-subject standard deviation units (SDw) using the SDw for each lung function test. The SDw is calculated from the difference between paired baseline measurements divided by 2; i.e., for sRaw: 0.109 kPa · s; Rint: 0.078 kPa · s · L¹; Xrs5: 0.104 kPa · s · L¹; and Rrs5: 0.131 kPa · s · L¹ (14). Thus the formula used in calculation of CACh response was:

(post CACh value  baseline value)/SDw

A change of 3 SDw was chosen as cutoff level for a definite response. This was used to calculate sensitivity, specificity, and predictive values of each lung function test. Paired t test was used to compare values before and after CACh. Unpaired t test was used to assess differences between groups. Repeatability was evaluated according to Bland and Altman (16) and by calculation of the correlation coefficient between first and second CACh. A p value of < 0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Thirty-eight asthmatic children were eligible for CACh test (Table 1). Their age ranged from 2.9 to 5.9 yr. The mean (± SD) period with a history of recurrent asthmatic symptoms was 18 ± 13.7 mo. Twenty-one (55%) had a first-degree relative with atopic disease and 18 (47%) were exposed to passive smoking at home. Twenty-two (58%) of the patients had other manifestations of atopy (atopic dermatitis and/or rhinitis) and 14 of 37 (38%) had at least one clinically relevant positive skin prick or RAST test. Until start of the observational period 34 (90%) of the patients were treated with inhaled corticosteroids, budesonide 100 to 1,200 µg/d (mean: 440 µg) delivered via a metered-dose inhaler and a metal spacer (Nebuchamber) or fluticasone 50 to 250 µg/d (mean: 133 µg) delivered via a plastic spacer (Babyhaler) and short acting ₂-agonist as required delivered via metered-dose inhaler and spacer device. The mean duration of the observational period, between coming off inhaled corticosteroids and qualifying for CACh, was 27 (range, 14 to 56) d. There was no significant correlation between duration of this period and the magnitude of responsiveness.

Thirty-three children age between 2.9 and 5.9 yr qualified as healthy control subjects. CACh was attempted in a total of 71 young children and completed in 67 (94%), 29 healthy and 38 asthmatic children. Four 2-yr-old children in the group of healthy children were not able to hyperventilate cold air at the desired level, although they did perform satisfactory lung function tests. All asthmatic children completed the test and CACh was well tolerated by all participants. One asthmatic child and four healthy control subjects were accompanied by a parent in the body box during measurements by whole body plethysmography. All other children accomplished the test while sitting alone in the body box.

Baseline measurements by each method are presented in Table 1 for both asthmatics and control subjects. The asthmatic children demonstrated statistically significant increased resistance at baseline values by all lung function measures in comparison to control subjects. Changes in lung function in each individual after CACh are shown in Figure 2. Hyperresponsiveness was detected by sRaw in 26 of 38 asthmatics versus 2 of 29 control subjects, by Rint in 12 of 38 asthmatics versus 1 of 29 control subjects, by Xrs5 in 9 of 38 asthmatics versus zero of 29 control subjects and by Rrs5 in 7 of 38 asthmatics versus 1 of 29 control subjects.

View larger version (21K): [in this window] [in a new window]   Figure 2.   CACh induced changes in lung function measurements (sRaw, Rint, Xrs5, and Rrs5) in asthmatics and healthy young children calculated as number of SDw: (post-CACh value  baseline value)/SDw. SDw designates the within-subject standard deviation. Means are indicated by horizontal lines. Dotted line indicates the 3 SDw change limit.

One child in the control group clearly demonstrated acute asthmatic reaction after provocation (Figure 2). This particular child had never shown symptoms of asthma, but nevertheless his response to CACh was significant by changes in both sRaw and Rint of 8 and 4 SDw units, respectively.

The group of asthmatics exhibited statistically significant mean (range) changes after CACh of 9.0 SDw units (0.9 to 32.1) in sRaw (p < 0.0001), 2.5 SDw units (2.2 to 13.0) in Rint (p < 0.05), 1.4 SDw units (2.4 to 8.7) in Rrs5 (p < 0.05), and 1.8 SDw units (1.5 to 14.0) in Xrs5 (p < 0.05) (Figure 1). However, there was a considerable overlap in response interval between asthmatics and healthy control subjects in both Rint, Rrs5, and Xrs5. Sixteen asthmatics (42%) exhibited clinical signs of acute asthma for a short period (1 to 6 min) after CACh; they all changed  8 SDw units in sRaw.

Sensitivity, specificity, and predictive values of each test in the diagnosis of hyperresponsiveness, defined as change > 3 SDw units, are given in Table 2. Overall, sRaw was the most useful parameter showing sensitivity of 68%, specificity of 93%, and predictive values of positive and negative test of 93% and 69%, respectively. Although the specificity of the other tests ranged from 97% to 100% all these parameters failed as useful tests by showing very low sensitivities ranging between 19% and 32%.

Thirteen of the 26 asthmatic children with a positive reaction to CACh measured by sRaw were scheduled for a second CACh test after 8 wk. The second CACh test was positive in 11 of 13 (85%), the correlation coefficient (95% confidence interval [CI]) was 96% (87 to 99%) (p < 0.0001). The Altman-Bland plots shown in Figure 3 demonstrate the agreement between first and second CACh response as measured by sRaw. No major bias was found. The other lung function measures exhibited poor repeatability (not shown).

View larger version (13K): [in this window] [in a new window]   Figure 3.   Repeatability of sRaw response to CACh within an interval of 8 wk, Altman-Bland plots.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There is a need for a simple test to separate hyperresponsive patients from normoresponsive patients within the heterogeneous group of young children with asthmatic symptoms. Hyperresponsiveness strengthen the asthma diagnosis, which is important for clinical practice and provides the means for more specific clinical trials of asthma treatment. This is the first study to demonstrate the feasibility of CACh in awake young children 2 to 5 yr of age. We demonstrated that it is possible to disclose bronchial hyperresponsiveness in awake young children and to separate asthmatics and healthy young children using CACh test as provocative stimulus and measurements of sRaw to quantify the bronchial response. Hyperventilation of cold, dry air has been shown to be a potent stimulus to bronchoconstriction in asthma and has been applied in many studies in both adults and schoolchildren (2).

Pharmacological tests of bronchial responsiveness, such as inhalation of histamine and methacholine have been extensively studied in children and also with success in young children of 2 to 5 yr (12, 13). However, these methods are notoriously time-consuming (8), implicating multiple-step protocols for construction of dose-response curves, and also imply risks of side effects. Furthermore, they are probably not mimicking the pathophysiology of asthma to the same extent as CACh (11). Exercise challenge performed as free running has been used in preschool children (17), but standardization aiming at constant submaximal workload is not always possible, increasing the risk of false-negative tests.

CACh is acceptable to patients and parents and may provide more clinically relevant information than pharmacological tests by evoking release of mediators from inflammatory cells instead of direct constriction of bronchial smooth muscle (11). An additional advantage may be avoidance of the problem of size correction involved in aerosol stimuli (11), though it has been argued that CACh does not standardize the temperature at each level of airway generation in different sized children (18).

The CACh technique used in this study is simple. Hyperventilation was achieved by having the child compete with a computer animation of a balloon. By using the face mask fitted with a mouthpiece during CACh we effectively prevented nose breathing, thereby optimizing the challenge stimulus of lower airways. We applied the single-step method because multiple-step protocols are more time-consuming and identify the same subjects as either normoresponsive or hyperresponsive (8, 11).

Previous studies (6, 7) and our own pilot study in 14 asthmatic schoolchildren showed that the optimal time interval for measuring maximal response is 3 to 5 min after CACh. To obtain measurements by all three methods during these 2 min it was only possible to perform duplicate measurements for sRaw and single determinations of Xrs5, Rrs5, and Rint. Response measurement was initiated with sRaw to allow for the time-consuming equilibration of the whole body plethysmograph (45 s) to take place in the period before lung function testing.

Because response to CACh is a very brief and transient phenomenon, in contrast to exercise-induced bronchoconstriction (4), we may have compromised the precision of Rint and IOS, which in fact could explain the higher sensitivity for sRaw in detecting bronchoconstriction after CACh. The sensitivity of CACh and sRaw measurements in this study (68%) is not quite comparable to sensitivities of 77 to 100% reported in studies on schoolchildren using CACh and spirometry (3, 7). First, this difference is probably explained by the uncertainty of the clinical asthma diagnosis in young children compared with schoolchildren. Sensitivity and specificity are critically dependent on how well asthma can be defined as well as how well we can rely on history from the healthy control subjects. Certainly, the history of both groups was equally carefully scrutinized, but with the limitation inherent to history taking in young children through their parents. The asthmatic children in this study were carefully selected based on both history of recurrent asthmatic episodes, need for current antiasthmatic therapy, and relapse of asthmatic symptoms during interruption of treatment with inhaled corticosteroids. Insufficient provocative stimulus may alternatively have reduced the rate of hyperresponsiveness. In schoolchildren and adults the appropriate level of voluntary hyperventilation to induce bronchoconstriction is usually calculated as 25 × FEV₁ per minute using the prechallenge FEV₁ measurement (7). Because young children can not perform reliable FEV₁ measurements, we chose to calculate the required hyperventilation rate from the body weight of the participants estimating that 1 L/min/kg of body weight approximates 25 × FEV₁ per minute. All children in this study accomplished hyperventilation at the required rate.

The repeatability over 8 wk of the sRaw response to CACh (r = 96%) was at least as good as the repeatability over 2 wk (r = 82%) reported in schoolchildren using spirometry (5).

CACh in young awake children is a relatively simple test but it depends on the individual child feeling comfortable with the procedure. The majority of children had fun during testing. No discomfort or complications were observed or reported. The pilot study of Ptc_CO2 monitoring during CACh showed that Ptc_CO2 rose by 0.3 kPa, which is without clinical significance. Since 1996 we have performed more than 400 CACh tests in young children without any complications. In those children disclosing a clinical reaction to CACh the duration was brief and transient without need for rescue medication.

Eucapnic voluntary hyperventilation using dry, room temperature air is a simplified technique, which in adult asthmatics demonstrated even higher sensitivity than CACh (19). The potentials of this technique for bronchoprovocation in young children should be evaluated in future studies.

The statistically significant difference in baseline lung function between asthmatics and healthy children indicates that even baseline measurements by these methods may help in differentiating healthy and asthmatic children.

In conclusion, this study demonstrates that CACh in young children is feasible and may serve as a useful tool in the study of bronchial responsiveness in asthmatic young children. The test is simple and imposes no discomfort on the child. Whole body plethysmography (sRaw) serves as the most sensitive and repeatable method in detecting change in lung function after CACh. Further studies are needed to define the usefulness of CACh and sRaw in the study of the effect of antiasthmatic medication in young asthmatic children.