INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Asthma is the most common cause of respiratory disability in children. It is recognized as causing significant morbidity and mortality (1). One of the common factors associated with asthma fatality is failure to recognize the severity of the asthma attack. This failure can be on the part of the physician, child, or family. It can result in an inappropriate, even fatal, delay in receiving treatment (2, 3). A critical patient factor in recognizing the severity of an asthma exacerbation is appropriate recognition of increased mechanical load, which should occur early in an asthma attack and induce the child to seek treatment. There may be two components in assessing the severity of asthma: an emotional or psychological component and a physiological component. Patients with asthma often deny the overall seriousness of their disease and therefore disregard their symptoms, resulting in a delay in seeking treatment. In addition, poor perception of their airway obstruction at a physiological level can result in an underestimation of the severity of the asthmatic attack, which can lead to a delayed response or complete disregard for the asthmatic symptoms.

In acute exacerbations of asthma, symptoms generally correlate well with impairment of lung function and patients are often adept in judging the degree of acute airway obstruction (4). However, it has been recognized that asymptomatic patients with chronic asthma may have severely reduced airway conductance (5). In addition, individuals with asthma who are seeking medical attention have been reported to have greater shortness of breath than hospitalized patients not claiming dyspnea and who have the same degree of functional impairment (8).

Psychophysical studies of respiratory load perception have been conducted in patients with asthma and normal subjects, using bronchial provocation (5, 7), external resistance (9), or both (13). In patients with asthma, there is significant variability in the perception of added loads that does not correlate with age or measures of lung dysfunction (14). One reason for this variability may be that there is a subpopulation of asthmatic patients with impaired ability to sense respiratory loads grouped with a much larger population of asthmatic patients with normal sensitivity. Kikuchi and coworkers (15) found that adult patients with life-threatening asthma have a reduced sensitivity to resistive loads as measured by a decreased slope in the resistive load-magnitude estimation curve. This is supported by a report from this laboratory that children with a history of life-threatening asthma also have a significantly decreased slope in their resistive load-magnitude estimation curve (16). Thus, it was hypothesized that there is a subpopulation of patients with asthma who have an impaired ability to sense mechanical loads, and this places these individuals at greatest risk of death from an acute asthmatic episode. This further suggests that these patients have a deficit in the neural processing system mediating respiratory load perception.

The fact that subjects are consciously aware of breathing against mechanical loads suggests that there must be activation of neurons in the cerebral cortex and this activation should be measurable. The activation of cortical neurons by mechanical loads has been studied by evoked potential techniques similar to those routinely used in other somatosensory systems. A mechanical load (inspiratory occlusion) applied while simultaneously recording from the somatosensory region of the cortex in the adult human (17) resulted in the observation of occlusion-elicited (respiratory-related) evoked potential (RREP). The first peak, P₁, was a positive voltage suggested to be due to the dipole that occurs when a cerebral cortical column was depolarized by the arrival of activity from a population of afferents that were activated by the occlusion stimulus. The P₁ peak may therefore signal the arrival of the inspiratory load related afferent information at the somatosensory cortex (18, 19). Subsequent studies in normal adult subjects have demonstrated that the P₁ peak of the RREP is elicited by resistive loads and the amplitude of this peak is correlated with the magnitude of the resistive load (20).

The present study applied the RREP technique to children with severe asthma who were identified as high risk for a life-threatening asthmatic attack. The children were classified as high risk on the basis of the criteria that these children often reported feeling asymptomatic when their measures of lung function indicated impairment and a history of asthma exacerbation severe enough to require hospitalization in an intensive care unit for respiratory failure (21). The RREP was recorded by inspiratory occlusion in three groups of children: (1) children with a history of life-threatening asthma (2) asthmatic children without a history of life-threatening attacks, and (3) nonasthmatic children. The results demonstrated a unique subpopulation of children with life-threatening asthma in which the P₁ peak was not present.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

All subjects, nonasthmatic and asthmatic, were recruited from the age group of 6-18 yr (three asthmatic subjects aged 21-23 yr were also studied). Nonasthmatic children (NA, n = 14) and asthmatic children without a history of life-threatening asthma (A, n = 15) were matched with the life-threatening asthma group (LTA, n = 11) for sex, race, and age to form the three experimental groups. NA subjects were classified as normal on the basis of habitual good health, no history of cardiorespiratory disease, no history of smoking, and no evidence of current major or minor illness. Children with asthma were recruited from the patient population at the University of Florida, Division of Pediatric Pulmonary Diseases. The diagnosis and rating of the severity of asthma were made according to standard clinical criteria. Two groups of children with asthma were tested: those with potentially life-threatening asthma (LTA) and those without (A). Criteria to identify children with life-threatening asthma were based on previous reports on the profile of patients who died from asthma (21). The subjects with LTA were asymptomatic without exacerbation of their asthma within 4 wk of the study. All these subjects had been admitted to the pediatric intensive care unit with acute respiratory failure, that is, with severe hypoxia and retention of carbon dioxide. The subjects with LTA were maintained on inhaled corticosteroids and theophylline. The subjects in the control A group were also receiving daily maintenance treatment with inhaled corticosteroids to control their symptoms. However, the control A subjects had never been admitted to the intensive care unit because of respiratory failure. The NA subjects were recruited from the local schools and were free of chronic respiratory disease. These subjects were also free of any acute respiratory disease at least 4 wk before the study. The experimental protocol was reviewed and approved by the University of Florida Institutional Review Board.

General Methods

The potential risks and benefits were discussed with the participants and their parent(s) or legal guardian(s). Once a patient met the criteria for inclusion in the study, assent was obtained from the subject and consent was obtained from the parents or guardians. The subjects were studied seated, at rest, with the back, neck, and head comfortably supported. Surface cup electrodes were placed at C_Z, C₃, and C₄, with the left ear as ground, to record scalp electroencephalogram (EEG) activity. The electrode positions were based on the international 10-20 system. The impedance level for each electrode was checked and maintained below 3 k. The electrodes were connected to an electroencephalograph system (Grass [Quincy, MA] model 12 neurodata acquisition system). EEG activity was recorded bilaterally with the vertex as reference, C_Z -C₃ and C_Z -C₄, and monitored with an oscilloscope (5111A; Tektronix, Beaverton, OR). The EEG activity was bandpass filtered (0.3 Hz-1 kHz), amplified, and led into an on-line signal averaging computer system (Cambridge Electronic Design [Cambridge, UK] model 1401, and IBM-compatible computer). Each computer sample of the analog traces was stored on computer disk.

The subject respired through a mouthpiece (Figure 1) and nonrebreathing valve (2600 series; Hans Rudolph, Kansas City, MO). The inspiratory port of the breathing valve was connected to an occlusion valve (Hans Rudolph series 9300). Mouth (Pmo) and occlusion (Pb, b = balloon) valve pressures were sensed with differential pressure transducers and signal conditioners. Pmo and Pb were displayed on the oscilloscope and led into the signal-averaging computer. Pb was used to trigger the collection of a 400-ms epoch of EEG and Pmo. The EEG activity and Pmo were digitized at 3 kHz.

View larger version (24K): [in this window] [in a new window]   Figure 1.   Schematic representation of the experimental preparation.

Protocol

The nonasthmatic subjects and patients with asthma were in good general health with no history of neurological disease or respiratory disease other than asthma. In addition, any patient with a history of exacerbation of his/her asthma within 4 wk of the study were excluded.

Each child was prepared for recording EEG activity as described above. They were seated in the chair and wore headphones connected to a television/video cassette player. The child watched a video tape throughout the experiment, with the sound masking experimental noises. This also significantly increased the duration the child tolerated the experimental procedure without attention-related interruptions. The child began respiring through the nonrebreathing valve, with the inspiratory port connected to the occlusion manifold (Figure 1). The inspiratory occlusion was presented by interruption of inspiration within 250 ms after the onset of the inspiratory effort. Each interruption occlusion was separated by two to six unoccluded breaths. The inspiratory occlusion was approximately 400 ms in duration. An individual occlusion trial consisted of 80 occlusion presentations. Control evoked potentials were obtained by turning a three-way valve to exclude the occlusion valve from the inspiratory circuit (Figure 1). The occlusion valve was again activated during inspiration but the subject's airway was not obstructed. The control trial consisted of 80 no-load presentations. A second occlusion trial was then presented. A 10-min interval separated each experimental trial with the child off the loading apparatus. Median nerve stimulation was applied to the patients with LTA as an additional somatosensory evoked potential control.

Data Analysis

The RREP for each trial was averaged separately (SIGAVG; Cambridge Electronics Design). The individual presentations for a trial were recalled from computer memory and displayed. The presentation was added to the average if it had a stable preocclusion baseline and was free of voltage changes in excess of 50 µV. For each trial, a minimum of 64 presentations were averaged. The averaged EEG and Pmo were obtained for the two occlusion trials and the control trial. The averaged RREP was specifically analyzed to determine P₁ peak latencies and amplitudes. The zero-time point for determination of the latencies (22) was set at the point of onset of occlusion, using the averaged Pmo signal (Figure 2A). The results were subjected to analysis for the effects of subject group, race, and sex. The two occlusion trials for each subject were averaged and again analyzed for RREP peak activity. The averaged control trial was subtracted from the occlusion trials as an additional control for artifacts and the result analyzed for P₁ latency and amplitude. Comparison between groups was initially performed by ANOVA and post hoc between-group analysis was performed with the Student-Newman-Keuls test. The significance criterion was set at p < 0.05. 

View larger version (17K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 2.   (A) Respiratory-related evoked potentials recorded from CZ -C3 and CZ -C4 for one subject in the nonasthmatic group. The RREP was elicited by inspiratory occlusion. The averaged EEG activity was control subtracted. The P1 peak is indicated by the arrow. (B) Averaged control trial respiratory-related evoked potentials recorded from CZ -C3 and CZ -C4 for the same nonasthmatic child as in (A). The valve was closed but the airway was not occluded.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The RREP was observed bilaterally in the somatosensory region (C_Z -C₃ and C_Z -C₄) with midinspiratory occlusions in all NA children, and a representative tracing of the average of 80 occlusions is shown in Figure 2A. The children were cooperative and tolerated the protocol with little discomfort. The RREP was absent when the occlusion valve was activated during inspiration but the airway not occluded (Figure 2B). The NA group mean P₁ peak latency and amplitude are summarized in Table 1.

The A children without a history of life-threatening attacks had moderate to severe asthma (Table 2). These A children were also cooperative subjects and had no adverse reaction to the protocol. The RREP was observed bilaterally (Figure 3) with inspiratory interruption occlusions and was absent on unoccluded breaths. The P₁ peak was present in 14 of 15 of these patients. The A group mean latency and amplitude are summarized in Table 1.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Respiratory-related evoked potentials recorded from CZ -C3 and CZ -C4 for one subject in the control asthmatic group. The RREP was elicited by inspiratory occlusion. The averaged EEG activity was control subtracted. The P1 peak is indicated by the arrow.

The patients with LTA had moderate to severe asthma (Table 3), similar to the control patients with asthma. There was no significant difference between the A and LTA groups for FVC, FEV₁ and %FEV₁/FVC. The P₁ peak was absent in the inspiratory occlusion trial in 6 of the 11 children (Figure 4A). The RREP was absent when the occlusion valve was activated during inspiration but the airway not occluded (Figure 4B). The somatosensory evoked potential elicited by median nerve stimulation was present in all these patients. The RREP was present bilaterally in four of these patients (Figure 4C). The LTA group mean P₁ latency and amplitude are summarized in Table 1. There was no significant difference in the P₁ latencies or amplitudes for all subject groups (NA, A, and LTA) when inspiratory occlusion elicited the P₁ peak of the RREP.

View larger version (17K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   Figure 4.   (A) Respiratory-related evoked potentials recorded from CZ- C3 and CZ -C4 elicited by inspiratory occlusion for one of the six subjects in the LTA asthmatic group where the P1 peak was not present. The averaged EEG activity was control subtracted. (B) Averaged control trial respiratory-related evoked potentials recorded from CZ -C3 and CZ -C4 for the same LTA asthmatic child as in (A). The valve was closed but the airway was not occluded. (C ) Respiratory-related evoked potentials recorded from CZ -C3 and CZ -C4 elicited by inspiratory occlusion for one of the five subjects in the LTA asthmatic group where the P1 peak was present. The averaged EEG activity was control subtracted. The P1 peak is indicated by the arrow.

Only one subject in the A group did not have the P₁ peak of the RREP (Figure 5). This patient was a male asthmatic patient with a history of noncompliance with medication, and who failed to keep appointments and was under consideration for admission into the program monitoring patients at high risk of a life-threatening asthmatic attack.

View larger version (14K): [in this window] [in a new window]   Figure 5.   Respiratory-related evoked potentials recorded from CZ -C3 and CZ -C4 elicited by inspiratory occlusion for the one subject in the control asthmatic group where the P1 peak was not present. The averaged EEG activity was control subtracted.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experiments performed to date in our laboratory have demonstrated that the RREP can be recorded in nonasthmatic and asthmatic children with a cephalic reference, C_Z. The significance of the present study lies in the identification of a subpopulation of children with a history of life-threatening asthma that do not have cortical neural activity normally elicited by an inspiratory load present in all other children tested.

The patients with LTA, when compared with the control children with asthma, were not significantly different in ventilatory status and asthma stability. All these patients with LTA, however, had been admitted to the pediatric intensive care unit with acute respiratory failure within 4 yr of the study. After their life-threatening event, the children with LTA were stabilized and maintained with inhaled steroids and theophylline. The subjects in the control A group, similar to the children with LTA, had moderate asthma and were receiving daily maintenance treatment (inhaled steroids) to control their asthma symptoms; however, they had never been admitted to the intensive care unit for respiratory failure. Thus, these two groups had asthma of similar severity which was under sufficient control for them all to be asymptomatic for at least 4 wk before study. The primary difference between these two groups of asthmatic children was the occurrence of a life-threatening asthmatic attack. The cause of the life-threatening event was not known but may be due, in part, to a delayed perception of symptoms by some of these children. Thus, the absence of the P₁ peak of the RREP in the children with LTA was not due to a difference in airway mechanics, suggesting a sensory processing deficit unique to these patients with LTA.

The P₁ peak of the RREP has been recorded with both a C_Z and joined earlobe reference (17, 22, 23). This peak is the result of a dipole that appears to be cortical in origin (19, 23). Respiratory mechanoreceptors in the respiratory tract and pump activated by the inspiratory load project to the central nervous system, eventually eliciting neural activity in the cerebral cortex. The depolarization of a population of cortical neurons produces a dipole that is recorded with surface electrodes as a voltage change. In a study of the distribution of the RREP recorded over the surface of the scalp, the P₁ peak was found bilaterally and of greatest amplitude when recorded over the somatosensory region. Webster and Colrain (19) have, in a detailed study of the cortical neural generators of the early peaks of the RREP, reported a P_1A peak that corresponds to the P₁ peak in the present study. Their P_1A peak was most likely produced by a dipole in the somatosensory region of the cerebral cortex. This means that the P₁ peak is a cortical marker of the arrival of the mechanosensory signal at the somatosensory region of the cerebral cortex. The absence of this peak in some children with LTA indicates that the occlusion-related mechanosensory signal is not activating these cortical neurons, i.e., the signal does not reach this region. This is further supported by the presence of a somatosensory evoked potential elicited by median nerve stimulation, indicating that this region of the cortex can be activated but not with respiratory mechanical stimulation. These results indicate that the children with LTA who do not have the P₁ peak have a deficit within the mechanoreceptor-to-cortex portion of the perceptual pathway. This does not preclude alternative pathways for respiratory sensation, which is a topic for future investigation.

Knafelc and Davenport (20) reported that the amplitude of the P₁ peak is correlated with the magnitude of inspiratory loads, the amplitude being greatest with large loads. This correlation of P₁ amplitude with load magnitude and magnitude estimation of the load (20) suggests that this cortical region is involved in the early processing of inspiratory load information. In the LTA children with an absent P₁ peak, this normal processing of inspiratory load information appears to be altered. This suggests that these children have a deficit in the cortical input pathway, which may contribute to an altered perception of the mechanical status of their respiratory system. While the P₁ peak is related to the arrival of mechanosensory information, later cognitive peaks, such as the P₃₀₀ (24- 26) may also be altered in these children. Future studies will be important to determine if these late cognitive-related peaks are present in these LTA children with an absent P₁ peak.

The early peaks of the RREP were present in the nonasthmatic and control asthmatic children and absent in 6 of 11 children with life-threatening asthma. The RREP has previously been elicited in adults with inspiratory occlusion, using a cephalic C_Z reference (17, 18, 22). The early peaks of the RREP have also been observed in children, using both a cephalic (C_Z) and noncephalic (joined earlobes) reference (23). In all these studies of nonasthmatic subjects, the early peaks were present. The absence of the early peaks of the RREP in the majority of the subjects with life-threatening asthma is unique to this patient group. The absence of the early peaks of the RREP in the children with life-threatening asthma cannot be the result of asthma, as evidenced by the presence of the early peaks of the RREP in the control asthmatic group. The absence of these early peaks of the RREP is also not due to differences in sex, race, or age. The children with life-threatening asthma appear to be a unique, small subpopulation of patients with asthma. The University of Florida maintains a list of children at high risk of a life-threatening asthma attack (21) and has had 20-40 children identified and included on this list. The fact that these children were already identified and separated into the LTA subpopulation made it possible to study this specific group with a majority having an altered or absent P₁ peak of the RREP. When all subjects with asthma are treated as a single test group, the small number of subjects with LTA is obscured by the overwhelming number of non-LTA patients. Thus, it is clear that the asthmatic population is heterogeneous; even the LTA subpopulation has group members that have the P₁ peak, have an altered P₁ peak, and do not have the P₁ peak. It is, however, important to recognize that there is a subpopulation of children with asthma that may be at high risk of a life-threatening asthmatic attack due to an altered neural processing (self-assessment) of their ventilatory status.

Psychophysical studies of respiratory load perception have been conducted in patients with asthma and normal subjects, using bronchial provocation (5, 7), external resistance (9), or both (13). Poor perception of asthma symptoms caused by bronchoprovocation was shown to occur in a subpopulation of patients with asthma in the seminal study by Rubinfeld and Pain (7). Kikuchi and coworkers (15) studied adults with near-fatal asthma and found a reduction in the perceptual sensitivity to resistive loads compared with normal controls (15). A similar-magnitude estimation study of children was performed in our laboratory (16). Five of the children with LTA in the present study were subjects in the study by Kifle, and coworkers (16). The P₁ peak was present bilaterally in one of these five subjects and unilaterally in one subject. The ME slopes for subjects with an absent or altered RREP were among the lowest for the LTA subject pool in the study by Kifle and coworkers (16). The mean slope for these four LTA subjects was 0.66 (mean slope for nonasthmatic subjects was 0.96), indicating a reduction in load perceptual acuity (16). This suggests that their reduced ability to sense the mechanical changes in their breathing associated with the onset of an asthmatic attack may be due to an altered neural processing of their ventilatory status and makes them at risk of a life-threatening event due to a reduced ability to self-assess their breathing.

In summary, the results of this study provide evidence that there is a unique subpopulation of children with life-threatening asthma who may have a difference in the neural processing of respiratory load information. The absence of the P₁ peak of the RREP in these patients suggests that the somatosensory cortex is not activated in the same manner as in other asthmatic and nonasthmatic children. Future studies will be necessary to determine the specific neural mechanisms responsible for these differences. The identification and enhanced monitoring of this subpopulation may aid in the prevention of future life-threatening events for these children.