Chest ACCP Member Benefits
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Holliday, J. E.
Right arrow Articles by Lippmann, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Holliday, J. E.
Right arrow Articles by Lippmann, M.
(Chest. 2003;124:1500-1511.)
© 2003 American College of Chest Physicians

Reduction in Ventilatory Response to CO2 With Relaxation Feedback During CO2 Rebreathing for Ventilator Patients*

Jerome E. Holliday, PhD and Michael Lippmann, MD, FCCP

* From the Veterans Affairs Medical Center, St. Louis MO.

Correspondence to: Jerome E. Holliday, PhD, Medicine Services, 151JC, VA Medical Center, 915 N Grand Blvd, St. Louis, MO 63106; e-mail: HollidayJEH{at}msn.com


   Abstract
TOP
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Background: Previous studies have shown that relaxation biofeedback reduced the amount of time spent receiving ventilation for difficult-to-wean patients.

Objective: The present study was begun to test the hypothesis that the underlying mechanism of biofeedback ventilator weaning was the reduction of neural respiratory drive (NRD).

Design: Prospective.

Setting: Pulmonary Medicine division in a Veterans Affairs hospital and the St. Louis Regional Medical Center.

Subjects: Twenty-four patients who were receiving mechanical ventilation were randomly assigned to either the biofeedback group or the control group.

Intervention: Respiratory relaxation feedback (RFB) was administered while a single variable, PaCO2, was inputted to the respiratory control system and the output was measured. While rebreathing 7% CO2/93% O2, the biofeedback group received a CO2 trial session and a CO2 RFB session, and the control group received a CO2 trial session and a CO2 no-feedback (NFB) session.

Measurements and results: There was a significant (p < 0.05) reduction in respiratory and EEG parameters for the RFB group at maximal end-tidal CO2 (mean [± SE], 70 ± 0.2 mm Hg) between the CO2 trial and the CO2 RFB session compared to the control group where there was no significant difference between the results of the CO2 trial and the CO2 NFB session. The mean values for the CO2 trial and CO2 RFB session for the biofeedback group were as follows: occlusion pressure 0.1 s from the onset of inspiration, 8.42 ± 1.08 and 6.48 ± 0.78 cm H2O (which reflects NRD), respectively; minute inspiratory ventilation, 15.84 ± 0.81 and 13.91 ± 0.72 L/min, respectively; mean inspiratory flow, 670 ± 2.28 and 581 ± 35 mL/s, respectively; respiration rate, 32 ± 2.28 and 31.2 ± 2.58 breaths/min, respectively; and chest background electromyography, 4.89 ± 0.71 and 3.54 ± 0.54 µV, respectively. The mean electroencephalograph outputs for the CO2 trial and CO2 RFB session for the biofeedback group were as follows: mean EEG frequency, 14.78 ± 0.98 and 13.06 ± 0.59 Hz, respectively; and beta EEG power, 3.1 ± 0.03 and 2.39 ± 0.19, µV2, respectively; and gamma EEG power, 2.96 ± 0.34 and 2.24 ± 0.24 µV2, respectively.

Conclusion: We conclude that the decrease in respiratory parameters reflecting NRD induced by RFB represents a key mechanism for the previously demonstrated effectiveness of biofeedback in reducing weaning time from mechanical ventilation.

Key Words: CO2 rebreathing • EEG • neural respiratory drive • ventilator • weaning


   Introduction
TOP
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Previous investigators1 2 3 have described the influence of respiratory drive on weaning from mechanical ventilation. These reports indicated that ventilator weaning was more difficult if neural respiratory drive (NRD), as measured by the change in endotracheal occlusion pressure 0.1 s from the onset of inspiration (P100),4 5 6 was > 4 cm H2O. In a randomized study, Holliday and Hyers7 showed that relaxation feedback reduced ventilator weaning time in a group of ventilator-dependent patients by 12 days. Their results suggested that the reduced time spent receiving ventilation might be the result of relaxation biofeedback decreasing the NRD. Holliday and Veremakis8 have shown that the relaxation feedback technique used in the ventilator study by Holliday and Hyers7 reduces NRD in healthy subjects undergoing hypercapnic challenge, as measured by inspired minute ventilation (I), P100, mean inspiratory flow (VT/TI), and respiration rate (RR). These parameters are common measures of respiratory9 output during CO2 rebreathing and are important respiratory parameters (RPs) in ventilator weaning in subjects undergoing hypercapnic challenge. Studies by Bulow,10 Wolkove and coworkers,11 and Asmussen12 have shown that drowsy and meditative states reduced the respiratory output of I during CO2 rebreathing compared to the normal waking state.

The theory of Shea13 states that activity in the reticular activating system (RAS), which affects the NRD, parallels the state of arousal. One indicator of increased arousal is an increase in EEG power, particularly in the beta and gamma bands (ie, 14 to 35 Hz). Conversely, a reduction in beta and gamma power indicates reduced activation. Crippen14 has shown that increased alertness and anxiety in the ICU patient due to the exhaustive ICU environment results in increased power in the EEG bands, especially in the beta band. Bulow10 showed that the stress of doing arithmetic problems increased the power in the beta band compared to a normal awake state and increased the I for a given end-tidal CO2 (ETCO2) during CO2 rebreathing in healthy subjects. Conversely, a reduction in beta and gamma power in the EEG would be an indication of reduced activation.

We hypothesized that there would be a significant reduction in the RPs of P100, (reflecting NRD), and I, VT/TI, and a reduction in EEG activation during CO2 rebreathing with relaxation feedback compared to CO2 rebreathing without relaxation feedback in intubated ICU patients prior to extubation. We investigated whether a reduction of NRD is the possible reason that biofeedback relaxation reduces weaning time from mechanical ventilation.

This article attempts to demonstrate the mechanism through which biofeedback reduces ventilator weaning time from mechanical ventilation, as shown in the study by Holliday and Hyers.7 Thus, the weaning time of individual patients receiving mechanical ventilation will not be considered.


   Materials and Methods
TOP
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Subjects
The study was conducted at the St. Louis Regional Medical Center and the John Cochran Veterans Affairs Medical Center. After the study was approved by the Institutional Review Board of the St. Louis Regional and Veterans Affairs Medical Centers, 12 patients receiving mechanical ventilation (9 men and 3 women), with a mean (± SE) age of 62.6 ± 4.03 years, were randomly assigned to the biofeedback group and 12 patients (10 men and 2 women), with a mean age of 64.2 ± 3.05 years, were randomly assigned to the control group. Analysis of variance (ANOVA) indicated that the age difference between the groups was not significant. One patient in the control group could not complete the second CO2 rebreathing session, and his data were not included in the data for the control group. The patient selection criteria were as follows: (1) duration of mechanical ventilation of > 7 days; (2) the patient was clinically stable and had been afebrile for 24 h; (3) no sedatives or narcotics had been administered for at least 8 h before the study; (4) no bronchospasm or diaphragmatic disease, neurologic injury or neurologic disease; (5) negative inspiratory force of > 20 cm H2O; and (6) 20-min T-tube trial in which PaCO2 did not rise > 9 mm Hg and the patient remained stable with an RR/tidal volume (VT) ratio of < 100. These criteria assured that the patients could tolerate the CO2 rebreathing sessions without having to be returned to mechanical ventilation during the test. Patients were eligible regardless of previous weaning attempts because this study did not address weaning time. The subjects were not informed as to the outcome, hypothesis, or expected results of the research. Each subject was informed that they would have two sessions of spontaneously rebreathing CO2 through their endotracheal tube and a rest period between each session to allow time for the computer to record data on a disk.

Respiratory Measurements
RPs measured during CO2 rebreathing included I, VT/TI, RR, and P100. Unlike I and VT/TI, which have been used as measures of respiratory output in previous studies of CO2 rebreathing,15 16 P100 reflects NRD independent of respiratory system flow resistance or compliance.4 17 Reproducible P100 data can be obtained on conscious humans if they are obtained in the first 0.1 s of inspiration,4 which is before the subject is aware of the occlusion of the airway. A diaphragmatic electromyogram (EMG) would be a more direct measure of NRD than P100, but, due to a number of patients in our ICU having abdominal dressings, it was not possible to measure diaphragmatic EMG with surface electrodes. In the present study, I, VT/TI, VT, and RR were measured for comparison with P100 and for comparison with previous work on CO2 rebreathing. A schematic drawing of the hypercapnic challenge breathing circuit18 showing the CO2 rebreathing bag, the RFB system, and the two-way valve with balloon for occluding the airway for P100 measurements4 5 6 (model 9326; Hans Rudolph; Kansas City, MO) is shown in Figure 1 . In order for the patient not to control P100 consciously, they must not hear the valve opening to inflate the balloon. To ensure this, the valve was placed in a soundproof container. The P100 was measured every 20 s during CO2 rebreathing, or approximately every third or fourth breath. VT was measured by integrating the output of a heated pneumotachograph (model 3700; Hans Rudolph), which was calibrated before each session. The output of the pneumotachograph and the occlusion pressure were converted into a voltage using a pressure transducer and carrier demodulators (Validyne Corp; Northridge, CA). The outputs then were fed into a computer that performed the integration of the pneumotachograph output and measured the endotracheal occlusion pressure. The endotracheal occlusion pressure and the endotracheal tube airflow were recorded on a computer (HP 3100 computer; Hewlett-Packard; Palo Alto, CA; and IBM; New York, NY) using an software program (Spectra Plus; Pioneer Hill Software; Poulsbo, WA). The computer program displayed a time series plot, which is a two-channel oscilloscope on the computer monitor. Recording flow through the endotracheal tube on the left channel and mouth air pressure on the right channel on the two-channel oscilloscope allowed the start of the P100 to be determined from a mouth air flow of 0. The mean P100 for a given ETCO2 was obtained by taking the P100 value from each patient’s P100 vs the ETCO2 linear regression curve at that ETCO2. The value of the ETCO2 was obtained from the computer printout of the value of ETCO2 for a breath number (ie, the number of breaths taken during the session). Besides ETCO2 and P100, the RPs measured during each hypercapnic challenge were VT, I, VT/TI, RR, heart rate (HR), and O2 saturation (model 3700 pulse oximeter; Datex-Ohmeda; Madison, WI).The I, VT, VT/TI, and RR were recorded and computed on a computer (Apple IIe; Apple Computer; Cupertino, CA).



View larger version (28K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1.. Schematic diagram showing the patient sensor and computer console (A), RFB headset (B), abdomen movement detector (C), and CO2 rebreathing system. The CO2 circuit shows the rebreathing bag, pneumotachograph, and the two-way valve and balloon for measuring P100.

 
CO2 Rebreathing
The ETCO2 was measured with an infrared monitor (model 125; Puritan-Bennett Datex; Los Angeles, CA), which was calibrated before each session. A 5-L rebreathing bag was used to ensure the rapid mixing of gas between the bag and lungs at the onset of rebreathing. The mean ETCO2 values for the S and end of the session (after 2.5 min of rebreathing CO2) were 50 ± 0.86 and 70 ± 1.2 mm Hg, respectively. The mean increase of 20 mm Hg in ETCO2 is considered to be adequate for stimulating respiratory drive.18

EEG Measurements
The EEG was measured by two gold-plated electrodes (Grass Instruments; West Warwick, RI), which were placed in bipolar fashion on each side of the occipital lobe O1 and O2 locations according to the International 10–20 Electrode Placement system. The surface of the skin was prepared with an abrasive solvent (NuPrep; DO Weaver & Co; Aurora CO). This assured that the resistance between electrodes was < 2,000 ohm for all subjects. The EEG was amplified by a preamplifier (model P-15; Grass Instruments) and a sound card (Tahiti; Turtle Beach; Yonkers, NY), which had a flat frequency response in the EEG region. The preamplifier frequency range was set between 0.1 and 300 Hz. Using the time series of the computer program (Pioneer Hill) the EEG is displayed in the usual fashion with the horizontal axis showing the frequency and the vertical axis showing the amplitude in µvolts. Using computerized fast Fourier transform (FFT) it is possible to obtain delta (0.5 to 3 Hz) associated with sleep, theta (4 to 7 Hz) in the drowsy state, alpha (8 to 13 Hz) with eyes closed, which is associated with the relaxed state, beta (14 to 24 Hz) and gamma (25 to 35 Hz) with increased alertness and anxiety, and mean power amplitudes as a function of time. Although it has been traditional to consider only the dominant EEG band associated with a particular state of consciousness or sleep state, computerized FFT analysis has shown that all the spectral bands are present to some extent at all levels of consciousness. In order to obtain a measure of the change in band power between the start of the run and the finish of the run (ie, the mean EEG band power of the first [S] and last [F] 20 s of the run), a mean was obtained of each band power for the first and last 20 s of the run. The mean EEG frequency (MEEG) is the root mean square integrated value from 0.5 to 35 Hz for every 0.75 s for the entire time of the run and is the same method as that used by Lennox et al.19 The MEEG was obtained to show the shift in EEG spectral frequency toward lower frequency with RFB.

Biofeedback
We utilized a respiratory relaxation feedback (RFB) system developed by Leuner20 (DHD Healthcare Crop; Baldwinsville, NY). The RFB system is a noncognitive diaphragmatic breathing instrument. It is based on a diaphragmatic breathing principle used by van Dixhoorn,21 who showed in a controlled study on postmyocardial infarction patients that breathing training can produce relaxation, as evidenced by a significant decrease in HR and RR compared, to a control group. van Dixhoorn21 had his experimental subjects reduce their breathing rate and increase the distention of their abdomen by placing their hand on their abdomen to monitor its motion. The RFB system reduces RR and increases diaphragmatic breathing to produce relaxation electronically by delivering a continuous stimulus and response with each breath. The system consists of a headpiece with lights (mask) and earphones over the ears, and an infrared sensor that detected the distention of the abdomen. With inhalation, the abdomen distends causing a sound to be heard in the earphones and two lights to come on in the headpiece. On exhalation, the abdominal muscles relax, which turns off the sound and the lights. The sensitivity of the RFB system was adjusted for each individual so that the intensity of sound and light was comfortably seen and heard, and went on during inhalation and off on exhalation. The RFB system requires less concentration than the electromyogram (EMG) relaxation feedback used by Holliday and Hyers,7 and thus would be more effective in producing relaxation in patients receiving mechanical ventilation who have difficulty concentrating. In order to emphasize breathing from the diaphragm and a slow RR, the subjects were told to make the sound and light as intense and spaced as far apart as possible. At the start of the RFB session, subjects were given 30 s of instructions stating that the light and sound and their slow deep breaths were relaxing them, in order to help the patient associate the light and sound with relaxation.

EMG Measurements
In order to determine whether the RFB system was actually relaxing the patient, the EMG was measured between the third and fourth upper intercostal muscle (IC), which was the same location as the EMG feedback for the ventilator weaning study of Holliday and Hyers7 and the CO2 rebreathing study of Holliday and Veremakis.8 Reductions in the EMG commensurate with those in the above studies using EMG feedback would indicate that RFB is equivalent to EMG feedback in producing relaxation. For the chest EMG to reflect anxiety, the relaxing and contraction rhythm of the IC muscles that occurs during breathing were subtracted from the total EMG of the IC and the remaining EMG was referred to as chest background (CHBK) EMG. Thus, the reported CHBK EMG was proportional to the background EMG muscle tension. The EMGs were measured with surface electrodes using an electromyograph (model M-57; J&J Co; Poulsbo, WA) that measures integrated EMG amplitude in µvolts. The electrodes were placed according to the method of Gross and coworkers.22

Protocol
When the ventilator patient met the weaning criteria but had not been extubated, the patient was removed from ventilator support and was placed on supplemental oxygen via a T-tube apparatus. After the subject became accustomed to breathing without ventilator support and ETCO2 had achieved a steady state, baseline measurements of P100, I, VT/TI, RR, and CHBK EMG were obtained over a 5-min period. This helped to determine whether the groups had similar respiratory outputs prior to CO2 rebreathing. Immediately after the baseline session, the subjects in each group received in random order two 2.5-min sessions of rebreathing 7% CO2/93% O2. The first CO2 rebreathing session served as a CO2 trial session for both groups. In the second CO2 rebreathing session, the biofeedback group was given the RFB mask with feedback of their abdominal breathing from the light and sound, and this was called the CO2 RFB session. The second CO2 rebreathing session for the control group was a no-feedback (NFB) session during which the mask was worn but they did not receive auditory or visual feedback, and this was referred to as the CO2 NFB session. There was continuous monitoring of O2 saturation, ECG, RR, and BP. Neither group was given sympathy, encouragement, or coaching.

Computer Analysis of Data
The methods employed below were the same as used by Holliday and Veremakis.8 All the data including the ETCO2 were continuously fed into the computer, from the DC analog output of the instrument with the exception of P100, which was measured with every third breath. The plotting of P100 as a function of ETCO2 has been described under respiratory measurements. The computer correlated the RP values as a function of ETCO2 on a breath-by-breath basis. The computer then plotted the values of the parameter as a function of ETCO2 at each breath. A regression line curve of all the values was generated for each parameter as a function of ETCO2 between 50 and 70 mm Hg for each individual subject. Curves for the RFB group for each RP were generated for each RP, as a function of ETCO2, by taking the mean values of that parameter obtained from each subject’s linear regression curve between ETCO2 values of 50 and 70 mm Hg at 4-mm Hg intervals. The slope for the RFB group for a given parameter was obtained from the sum of the slope of each subject’s regression line curve for a given RP. In CO2 rebreathing studies, the effect of the inputted factor such as relaxation, meditation, or stress, is determine by a significant change in the RPs at the maximum value of ETCO2 (mean, 70 ± 1.2 mm Hg) and significant change in {Delta}RP/{Delta}ETCO2 slope.9 10

Statistical Analysis
The differences among the baseline respiratory and EEG measurements of the two groups, the respiratory and EEG measurements of the CO2 trial and the CO2 RFB session in the biofeedback group, and the CO2 trial and CO2 NFB sessions in the control group were compared for significance with a two-factor repeated measures ANOVA. All data are expressed as the mean ± SE, and all differences were considered to be significant at p < 0.05.


   Results
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Respiratory Output
Patient diagnosis is shown in Table 1 . The RFB and control groups were closely matched in regard to underlying disease and the reason for ventilator dependence. There were no significant differences in respiratory output parameters between the two groups during the baseline period of spontaneous breathing on a T-tube prior to the onset of CO2 rebreathing (Table 2 ). The baseline values in Table 2 give the normal values for the RPs. Figure 2 and Table 3 show the responses to CO2 rebreathing. At an ETCO2 of 70 ± 1.2 mm Hg for the CO2 trial sessions, both groups showed significant increases over baseline values (Tables 2 and 3) of all RPs and measures reflecting NRD. There was no significance difference in RPs for the CO2 trial sessions between the two groups (Table 3) .


View this table:
[in this window]
[in a new window]

  		Table 1.. Underlying Disease for ICU Admission*

 

View this table:
[in this window]
[in a new window]

  		Table 2.. Mean Values for 5-Minute Baseline Measurements Prior to CO2 Rebreathing*

 


View larger version (22K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2.. A: CHBK EMG; B: I; C: P100; D: VT/TI as a function of ETCO2 for trial and RFB during CO2 rebreathing. The points are mean values of the 12 RFB subjects that were obtained from linear regression plots of each subject ("Computer Analysis of Data" section).

 

View this table:
[in this window]
[in a new window]

  		Table 3.. Mean Values for CO2 RFB Compared to CO2 Trial for RFB Group and CO2 NFB Compared to CO2 Trial for the Control Group at ETCO2 70 mm Hg*

 
For the RFB group, during CO2 RFB, Figure 2 (top right, B, bottom left, C, and bottom right, D) and Table 3 show a significant (p < 0.05) reduction in the mean values at maximum ETCO2 for P100 of 1.94 ± 0.91 cm H2O, for I of 1.93 ± 0.76 L/min, for VT/TI of 89 ± 38 mL/s, and for RR of 1.7 ± 0.79 breaths/min, when compared to the initial CO2 trial. Figure 2 , top left, and Table 3 demonstrate that RFB significantly decreases (p < 0.05) CHBK EMG during mean CO2 rebreathing of 1.35 ± 0.62 µV (baseline normal value, 3.29 µV) at an ETCO2 of 70 ± 1.2 mm Hg compared to the CO2 trial for the RFB group. The decrease in CHBK EMG relative to the CO2 trial continued despite the lack of verbal instructions after an ETCO2 of 54 mm Hg and the significant (p < 0.05) reduction in CHBK EMG for the RFB session at an ETCO2 of 66 mm Hg indicated that RFB had led to the decrease. No significant change occurred for the control group between the CO2 trial and the CO2 NFB sessions in the above parameters at maximum ETCO2 (Table 3) . The lack of any significant change for CO2 NFB shows that there was no decrement in CO2 responsiveness the second time it was administered. In other words, there was no practice effect. This shows that the above decrease in RPs for CO2 RFB is not due to reduced responsiveness to CO2 the second time it is rebreathed.

The slope reductions for P100, I, and VT/TI as a function of ETCO2 for the CO2 trial compared to the CO2 RFB session are shown in Figure 2 (top right, bottom left, and bottom right) and Table 4 . (The method of obtaining the curves was described in "Computer Analysis" and for P100 in the "Respiratory Measurements" subsections of the "Materials and Methods" section.) There was a significant (p < 0.05) reduction in the slope of {Delta}I/{Delta}ETCO2, {Delta}P100/{Delta}ETCO2, and {Delta}VT/TI/{Delta}ETCO2 between the CO2 trial and the CO2 RFB session. No significant differences appeared in the mean slopes of these breathing parameters between the CO2 trial and the CO2 NFB sessions for the control group.


View this table:
[in this window]
[in a new window]

  		Table 4.. Mean Slopes for CO2 Trial Compared to CO2 RFB Session for RFB Group and CO2 Trial Compared to CO2 NFB Session for the Control Group*

 
EEG Spectrum
The MEEG and spectral analysis of the EEG spectrum for the delta, theta, alpha, beta, and gamma bands is shown in Tables 5 and 6 using FFT. There were no significant differences between groups in the baseline value EEG bands at the S of the baseline period of breathing on a T-tube (Table 5 ; S-RFB vs S-control). There was also no significant difference in their final values shown by the p values for F-RFB vs F-control. In both groups, the delta band had the highest power and the gamma band had the lowest. Both groups demonstrated a significant (p < 0.045) increase in beta and gamma power during the 5 min of breathing room air and supplemental oxygen on a T-tube (Table 5 ; S-base vs F-base). The above S baseline values (Table 5) give the normal values for these EEG parameters.


View this table:
[in this window]
[in a new window]

  		Table 5.. Mean Relative EEG Band Power S and F Baselines Without CO2 Comparing Baselines of RFB and Control Groups*

 

View this table:
[in this window]
[in a new window]

  		Table 6.. MEEG and Mean Relative EEG Band Power Comparing F Between CO2 Trial and CO2 RFB Session and CO2 Trial and CO2 NFB Session at ETCO2 = 70 mm Hg*

 
From Table 6 , there was a significant (p < 0.05) reduction in MEEG frequency of 1.72 ± 0.74 Hz (CO2 RFB session, 13.06 ± 0.98 Hz, compared to the CO2 trial, 14.78 ± 0.59 Hz) at an ETCO2 of 70 ± 1.2 mm Hg, but not between CO2 NFB and the CO2 trial for the control group. A significant (p < 0.031) decrease of 0.71 ± 0.15 µV2 occurred for the CO2 RFB session (2.39 ± 0.31 µV2) compared to that for the CO2 trial (3.10 ± 0.26 µV2) for the beta band, and for the gamma band there was a significant (p < 0.017) decrease of 0.72 ± 0.32 µV2 for the CO2 RFB session (2.24 ± 0.34 µV2) compared to the CO2 trial (2.96 ± 0.29) at an ETCO2 of 70 ± 1.2 mm Hg. There were no significant changes for delta, theta, and alpha bands for the RFB group. There were no significant changes for any of the bands between the CO2 trial and the CO2 NFB session for the control group. Figure 3 shows a representative EEG tracing from one subject at an ETCO2 of 70 ± 1.2 mm Hg. The top tracing in Figure 3 is from the CO2 trial, while the bottom tracing is from the CO2 RFB session and shows the reduction in the beta and gamma EEG bands.



View larger version (30K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3.. Representative EEG tracings from one ventilator patient at the end of CO2 rebreathing for CO2 trial (top) and CO2 RFB session (bottom). A greater predominance of high-frequency beta and gamma EEG bands can be seen in the upper tracing, showing that RFB reduces the amount of high frequency in the EEG spectrum.

 

   Discussion
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Several studies7 23 24 25 26 27 have demonstrated that biofeedback facilitates the extubation of patients requiring mechanical ventilation. The purpose of this study was to demonstrate the possible mechanisms for the effectiveness of biofeedback relaxation in reducing weaning time. We demonstrated that the RFB technique used in this study improves RPs associated with successful extubation and reduces P100, a marker of NRD. Holliday and coworkers23 reported significant reductions in P100 and RR for high-drive ventilator patients between the start of weaning and successful extubation using EMG relaxation feedback. As noted, several investigators1 2 3 have reported that weaning was not successful when P100 was > 4 cm H2O. The significant reductions in P100 and RR during the CO2 RFB session (Table 3) likely contributed to the reduction in weaning time demonstrated by Holliday and Hyers.7

Montgomery et al28 showed that the mean value of the P100 for the patients failing weaning was 5.7 cm H2O compared to 3.7 cm H2O for the successfully weaned patients. However, they could not predict individual patient weaning failure from the measured P100 on ventilator patients during CO2 rebreathing. Their mean baseline P100 value of 4.7 cm H2O for ventilator patients is slightly higher than the present mean baseline value of 3.92 ± 0.84 cm H2O for both ventilator groups (Table 2) . However, our mean P100 of 8.18 ± 1.36 cm H2O at 70 mm Hg for the CO2 trials of both the RFB and control groups (Table 3) is higher than that found by Montgomery et al28 (6.8 cm H2O) during CO2 rebreathing, because their maximum CO2 value was 50 mm Hg. Yang and Tobin29 showed that an RR/VT ratio of < 100 predicts a high likelihood of weaning, whereas an RR/VT ratio of > 100 indicates a poor chance of weaning from mechanical ventilation. In a subsequent study of COPD ventilator patients, Jubran and Tobin30 found that the combination of increased mechanical load with rapid shallow breathing, leading to inefficient CO2 clearance, was the dominant mechanism of weaning failure. Although they did not attribute rapid shallow breathing to high NRD, they did state that reduced respiratory muscle activation would facilitate weaning from mechanical ventilation.

We have demonstrated that RFB can significantly (p < 0.05) reduce the CHBK EMG from 4.89 ± 0.71 µV (CO2 trial) to 3.54 ± 0.54 µV (CO2 RFB session) at an ETCO2 of 70 mm Hg, which is a reduction of 1.35 ± 0.62 µV (Table 3) . This is a greater reduction than the maximum reduction in CHBK EMG of 0.89 µvolts reported by Holliday and Hyers7 for EMG relaxation feedback, showing that RFB is as effective or more effective in producing relaxation as EMG relaxation feedback, even though the CO2 rebreathing is considerably more stressful than the weaning trials used by Holliday and Hyers.7 RFB also significantly reduced the subject values of P100, MEEG, I, VT/TI, and RR (Table 3) , which supports previous findings in healthy subjects that sleep or drowsy and meditative states decrease the RP response to CO2 during CO2 rebreathing,10 11 31 32 33 34 when compared to CO2 rebreathing in a normal awake state. However, this has not been previously demonstrated for patients receiving mechanical ventilation. The data in Table 3 show that during the CO2 trial session of the RFB group there was a greater increase in RPs than for the CO2 trial session of the control group, despite their having similar baseline values. The differences are not significantly different (Table 3) and likely represent individual variations in CO2 response. We therefore compared the values obtained during the CO2 trial with the CO2 RFB session of the RFB group to accurately reflect the effects of biofeedback relaxation. The RPs of the CO2 RFB response curves in Figure 2 are shifted to the right, and there is a flattening of the slope similar to those seen in previous studies.10 11 The 12% reduction in the mean value of I for RFB (Table 3) is less than the reduction reported by Bulow10 for the drowsy state and is similar to the reductions reported by Wolkove and coworkers11 for the meditative state in healthy subjects. This is not surprising given the stressful environment14 of the ICU, which makes relaxation difficult. The significant reduction in RR, but not in VT, shows that the significant reduction in I is due to a reduction in RR and not due to a reduction in VT. Brewis et al35 have stated that a >= 15% change in lung function is considered to be clinically significant. Thus, the 23% reduction in P100 is clinically significant, and the 12% reduction in I is close to clinical significance. The control group showed no significant change in parameters reflecting the NRD between the CO2 trial and CO2 NFB session. These results support the idea that biofeedback can reduce parameters reflecting NRD in patients who are receiving mechanical ventilation.

The changes in RPs and muscle relaxation were associated with RFB-induced EEG changes. Veselis et al36 showed that changes in the beta and gamma bands during sedation with midazolam were inversely proportional to the level of sedation for patients receiving long-term mechanical ventilation as measured by the Ramsay scale.37 The delta/beta power ratio for their subjects ranged from 2.22 at the lowest level of sedation at which patients could follow verbal commands to 12.5 at the highest level of sedation. In the present study, the baseline delta/beta power ratio was 1.56, indicating that our patients were alert and under little sedation. The delta power was the highest of the EEG bands for the patients receiving mechanical ventilation but is the lowest in awake healthy subjects. In the study by Veselis et al,36 the delta power was the highest for all of the EEG bands even for the lightest sedation, at which point the patients receiving mechanical ventilation responded to verbal commands. Kalkman et al38 showed that anesthetized patients have a delta/beta of 40.5, giving further evidence that the delta power level in the present research was well below that of sleep.

The fact that beta and gamma power decreases with increasing sedation indicates that these bands are affected by anxiety. It was indicated above that Crippen14 has shown that increased alertness and anxiety in the ICU patient, due to the exhaustive ICU environment, results in increased power throughout all the EEG bands, especially in the beta band. Table 5 shows that there was a significant (p < 0.05) mean increase in gamma power (0.32 ± 0.25 µV2) and beta power (0.24 ± 0.22µV2) between the S and F baselines for both groups while breathing supplemental oxygen prior to CO2 rebreathing, indicting that the patients receiving mechanical ventilation become more anxious the longer they are breathing without the aid of the ventilator. There is a significant (p < 0.031) reduction in beta and gamma power (Table 6) at an ETCO2 of 70 mm Hg for the CO2 RFB session relative to CO2 trial values, which is associated with the significant (p < 0.03) reduction in CHBK EMG (Table 3) . The significant reduction in beta and gamma power (Table 6) and CHBK EMG (Table 3) at an ETCO2 of 70 ± 1.2 mm Hg shows reduced anxiety for the ventilator patient, resulting in a significant (p < 0.05) reduction of P100, I, VT/TI, and RR, reflecting reduced NRD. The shift in the EEG spectrum toward lower frequency as seen by the significant (p < 0.05) decrease in MEEG shown in Table 6 at 70 ± 1.2 mm Hg appears to be due to the significant reduction in beta and gamma power due to RFB. There was no significant reduction in beta and gamma power or change in MEEG for the control group between the CO2 trial and the CO2 NFB session for an ETCO2 of 70 ± 1.2 mm Hg.

Shea13 has stated that the RAS has significant effects on the brainstem respiratory complex. Activity in the RAS increases with arousal and can provide stimulatory effects on certain brainstem-related neurons. Shea13 pointed out that signs of increased arousal include an increase in sympathetic nervous system activity and EEG activation, particularly an increase in power in the beta and gamma bands. Sensory stimulation is accompanied by increases in I, RR, and EEG activation14 and by slight increases in HR and BP. Bulow10 demonstrated an increase in the ventilatory response to CO2 rebreathing among subjects doing mental arithmetic. This increase was associated with an increase in beta power. The reduction in beta power during CO2 rebreathing in patients receiving RFB in this study (Table 6) indicates a decrease in RAS activation. A great amount of research has been done to show that muscle fatigue is an important factor in ventilator-weaning failure. However, little attention has been paid to the influence of the brainstem and RAS on setting NRD and its effect on ventilator weaning.

In summary, our study demonstrates that RFB during the stress of CO2 rebreathing leads to improvements in RPs associated with successful extubation. The above results show that RFB can reduce breathing parameters that reflect NRD for patients receiving mechanical ventilation. The study shows that the patient receiving mechanical ventilation has high anxiety when breathing without the help of the ventilator, which the patient has learned to rely on, increasing the NRD and making weaning difficult. However, biofeedback relaxation can improve weaning by reducing anxiety and thus NRD. This is the possible underlying mechanism for biofeedback reducing ventilator-weaning time in study by Holliday and Hyers.7 In addition, more research is needed on improving biofeedback techniques that further facilitate relaxation in patients receiving mechanical ventilation.


   Acknowledgements
 
The authors thank the Pulmonary Medical Staff of the Veterans Affairs Medical Center and the St. Louis Regional Medical Center for their cooperation and financial support during the study. A special thanks goes to the nursing staff of The ICU for assistance in performing the research.


   Footnotes
 
Abbreviations: ANOVA = analysis of variance; CHBK = chest background; EMG = electromyogram; ETCO2 = end-tidal CO2; F = mean EEG band power of last 20 s of run; FFT = fast Fourier transform; HR = heart rate; IC = intercostal muscle; MEEG = mean EEG frequency; NFB = no feedback; NRD = neural respiratory drive; P100 = occlusion pressure 0.1 s from the onset of inspiration; RAS = reticular activating system; RFB = respiratory relaxation feedback; RP = respiratory parameter; RR = respiration rate; S = mean EEG band power for first 20 s of run; I = inspired minute ventilation; VT = tidal volume; VT/TI = mean inspiratory flow

Received for publication March 4, 2002. Accepted for publication March 26, 2003.


   References
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Sassoon, CSH, Teresita, TT, Mahutte, CK, et al (1987) Airway occlusion pressure: an important indicator of successful weaning in-patients with chronic obstructive pulmonary disease. Am Rev Respir Dis 135,107-113[ISI][Medline]
  2. Herrera, M, Blasco, J, Venegas, J, et al Mouth occlusion pressure (PO.1) in acute respiratory failure Intensive Care Med 1985;11,134-139[ISI][Medline]
  3. Soma, K, Otsuka, H, Tomita, T Mouth occlusion pressure as a useful indicator for weaning from mechanical ventilation. Tohoku J Exp Med 1988;156(suppl),181-187
  4. Whitelaw, WA, Derenne, P, Milic-Emili, J Occlusion pressure as a measure of respiratory center output in conscious man. Respir Physiol 1975;23,181-199[CrossRef][ISI][Medline]
  5. Derenne, JP, Couture, J, Iscoe, S, et al Occlusion pressures in men rebreathing CO2 under methoxyflurane. J Appl Physiol 1976;40,805-814[Abstract/Free Full Text]
  6. Whitelaw, WA, Drenne, JP Airway occlusion pressure. J Appl Physiol 1993;74,1475-1483[Abstract/Free Full Text]
  7. Holliday, JE, Hyers, TM The reduction of weaning time from mechanical ventilation using tidal volume and relaxation biofeedback. Am Rev Respir Dis 1990;141,1214-1220[ISI][Medline]
  8. Holliday, JE, Veremakis, C Reduction in ventilator response to CO2 with relaxation feedback during CO2 rebreathing in normal adults. Chest 1999;115,1285-1292[Abstract/Free Full Text]
  9. Milic-Emili, J, Whitelaw, A, Grassino, A Measurements and testing of respiratory drive. Hornbein, T eds. Regulation of breathing: part III 1981,675-743 Marcel Dekker. New York, NY:
  10. Bulow, K Respiration and wakefulness in man. Acta Physiol Scand Suppl 1963;209,1-110
  11. Wolkove, N, Kreisman, H, Darragh, D, et al Effect of transcendental meditation on breathing and respiratory control. J Appl Physiol 1984;56,607-612[Abstract/Free Full Text]
  12. Asmussen, A Regulation of respiration the black box. Acta Physiol Scand 997;99,85-90
  13. Shea, SA Behavioral and arousal related influences on breathing in humans. Exp Physiol 1996;81,1-26[Abstract]
  14. Crippen, DW Neurologic monitoring in the intensive care unit. New Horiz 1994;2,107-120[Medline]
  15. Haldane, JS, Priestley, JG The regulation of the lung ventilation. J Physiol 1905;32,225
  16. Rebuck, AS, Slutsky, AS Measurement of ventilatory responses to hypercapnia and hypoxia. Hornbein, T eds. Regulation of breathing: part III 1981,745-771 Marcel Dekker. New York, NY:
  17. Grunstein, MM, Younes, M, Milic-Emili, J Control of tidal volume and respiratory frequency in anesthetized cats. J Appl Physiol 1973;39,395-404
  18. Read, DC A clinical method for assessing the ventilation response to carbon dioxide. Australas Ann Med 1967;16,20-32[ISI][Medline]
  19. Lennox, WG, Gibbs, FA, Gibbs, EL The relationship of cerebral activity to blood flow and to blood constituents. J Neurol Psychiat 1938;1,211
  20. Leuner, PL Oygraphic investigation of a respiratory feedback system applied on human beings [dissertation]. 1973 University of Gottingen. Gottingen, Germany:
  21. van Dixhoorn, J Cardiorespiratory effects of breathing and relaxation instructions in myocardial infarction patients. Biol Psychol 1998;48,123-135
  22. Gross, D, Grassino, A, Ross, WD, et al Electromyogram pattern of diaphragmatic fatigue. J Appl Physiol 1979;46,1-7[Abstract/Free Full Text]
  23. Holliday, JE, Shapiro, MJ, Durham, RM Optimization of P100 for reducing ventilator weaning time using tidal volume and relaxation feedback [abstract]. Am Rev Respir Dis 1991;143,A684
  24. Yarnal, JR, Herrell, DW, Sivak, ED Routine use of biofeedback in weaning patients from mechanical ventilation [letter]. Chest 1981;79,127
  25. Corson, JA, Grant, JL, Moulton, DP, et al Use of biofeedback in weaning patients from respirators. Chest 1979;76,543-545[Abstract/Free Full Text]
  26. Acosta, F Biofeedback and progressive relaxation in weaning the anxious patient from the ventilator: a brief report. Heart Lung 1988;17,299-301[ISI][Medline]
  27. Criner, GJ, Isaac, L Psychological issues in the ventilator-dependent patient. Respir Care 1995;40,855-865
  28. Montgomery, AB, Neagley, SR, Schoene, RB Prediction of successful ventilator weaning. Chest 1987;91,496-499[Abstract/Free Full Text]
  29. Yang, KL, Tobin, MJ A prospective study of indexes of predicting the outcome of trials of weaning from mechanical ventilation. N Engl J Med 1991;324,1445-1450[Abstract]
  30. Jubran, A, Tobin, MJ Pathophysiologic basis of acute respiratory distress in patients who fail a trail of weaning from mechanical ventilation. Am J Respir Crit Care Med 1997;155,906-915[Abstract]
  31. Reed, DJ, Kellogg, RH Changes in response to CO2 during natural sleep and at altitude. J Appl Physiol 1958;13,325-330[Abstract/Free Full Text]
  32. Bellville, JW, Howland, WS, Seed, JS, et al The effect of sleep on the respiratory response to carbon dioxide. Anesthesiology 1959;20,628-634[ISI][Medline]
  33. Gothe, B, Altose, MD, Goldman, MD, et al Effect of quiet sleep on resting and CO2-stimulated breathing in humans. J Appl Physiol 1981;50,724-730[Abstract/Free Full Text]
  34. Sautegeau, A, Hannhart, B, Peslin, R, et al Comparison between ventilatory and mouth occlusion pressure to hypoxia and hypercapnia in healthy sleeping man. Clin Physiol 1986;6,589-601[ISI][Medline]
  35. Brewis, RAL, Corrin, B, Geddes, DM, et al Respiratory medicine 2nd ed. 1955 WB Saunders. London, UK:
  36. Veselis, RA, Reinsel, R, Marino, P, et al The effects of midazolam on the EEG during sedation of critically ill patients. Anaesthesia 1993;48,463-470[ISI][Medline]
  37. Ramsay, MAE, Savege, TM, Simpson, BRJ Controlled sedation with apphaxolone-alphadolone. BMJ 1974;2,656-659[ISI][Medline]
  38. Kalkman, CJ, Boezeman, EH, Ribberink, AA, et al Influence of changes in arterial carbon dioxide tension on the electroencephalogram and posterior tidal nerve somatosensory cortical evoked potentials during alfentanil/nitrous oxide anesthesia. Anesthesiology 1991;75,68-74[CrossRef][ISI][Medline]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Holliday, J. E.
Right arrow Articles by Lippmann, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Holliday, J. E.
Right arrow Articles by Lippmann, M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS