INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

INSPIRATORY PRESSURE support ventilation (IPSV) is a popular form of partial ventilation that can be used in spontaneously breathing patients in a variety of clinical situations, including weaning from mechanical ventilation (16) and noninvasive ventilatory support in acute (5) and chronic (30) respiratory failure. During IPSV, each spontaneous breath is assisted by a constant level of positive pressure applied throughout inspiration. As a result, when ineffective effort and double triggering are not observed, breathing frequency is determined by the patient, and tidal volume (VT) depends on the combined action of the pressure generated by the inspiratory muscles, the ventilator, and the impedance of the respiratory system (25).

IPSV has been shown to increase alveolar ventilation and to reduce inspiratory effort (2, 12, 17, 25). The mechanism of inspiratory muscle inhibition during IPSV is not clear. A reduction in arterial PCO₂ (Pa_CO2) may be an important cause of respiratory muscle inhibition during IPSV. However, IPSV may also have a substantial inhibitory effect on respiratory activity even in the absence of Pa_CO2 changes. Two studies that compared inspiratory activity at the same end-tidal PCO₂ (PET_CO2) with and without IPSV found evidence of nonchemical inhibition of inspiratory activity during IPSV. Shams and Scheid (26) reported that anesthetized cats exhibited inhibition of inspiratory activity during IPSV, which persisted in part after correction of hypocapnia but was abolished when correction of hypocapnia was combined with vagotomy. However, these findings may not be generalizable to humans because vagal influences in humans are fairly weak. Scheid et al. (25) evaluated respiratory responses to inhaled CO₂ in normal human subjects during IPSV and found that IPSV effectively increased ventilation and reduced inspiratory activity at any given PET_CO2. However, they assessed respiratory center output on the basis of the occlusion pressure P_0.1, which only reflects the initial part of the respiratory drive and is only minimally informative as to what happens later during inspiration.

We conducted a study in healthy humans during IPSV in which 1) the PET_CO2 level was controlled by addition of CO₂ to the inspired circuit and 2) respiratory motor output was evaluated by recording esophageal (Pes) and transdiaphragmatic pressures (Pdi) and central respiratory output by recording the electrical activity of the diaphragm. Our objective was to determine the relative contributions of reduced PET_CO2 and mechanical unloading to inspiratory activity inhibition during IPSV. Specifically, we compared the parameters of inspiratory activity during spontaneous ventilation and IPSV before and after returning PET_CO2 to its spontaneous level by adding CO₂ to the inspired air during IPSV.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Glossary

CMV	Controlled mechanical ventilation
EMGdi	Electromyographic activity of the diaphragm
EMGdi,mTI	Inspiratory-integrated EMGdi
EMGdipeak	Peak inspiratory amplitude of the moving-time average of the EMGdi signal
EMGdi,Ttot	Total integrated EMGdi
IPS	Inspiratory pressure support
IPSV	Inspiratory pressure support ventilation
mTE	Mechanical expiratory time
mTI	Mechanical inspiratory time
P0.1	Occlusion pressure
PaCO2	Arterial PCO2
PAV	Proportional assist ventilation
Pdi250	Positive deflection of transdiaphragmatic pressure 250 ms after the onset of inspiratory activity
Pdi500	Positive deflection of transdiaphragmatic pressure 250 ms after the onset of inspiratory activity
PEEPi	Intrinsic positive end-expiratory pressure
PETCO2	End-tidal PCO2
PTPdi	Transdiaphragmatic pressure-time product
RR	Respiratory rate
Ttot	Total cycle time
E	Minute ventilation
VT	Tidal volume
	Experiments were performed in nine healthy volunteers (3 women and 6 men; age, 35 ± 5 yr; weight, 73 ± 8 kg; height, 173 ± 3 cm).

Measurements

All signals were sampled and digitized at 128 Hz, and the data were entered into a microcomputer by using an analog-digital system (MP100, Biopac System). In addition, EMGdi and flow signals were digitized at 1,000 Hz and sampled in a second microcomputer for subsequent analysis by using another analog-digital system (MP100, Biopac System).

Experimental Protocol

Data Analysis

The following variables were read breath by breath: mechanical inspiratory time (mTI), as the onset of the inspiratory flow to the onset of the expiratory flow; mechanical expiratory time (mTE), as the remainder of the total breath duration; VT from the calibrated integrated flow signal; and PET_CO2 as the peak of the airway CO₂ record. From these, the following parameters were calculated breath by breath: total respiratory time, T_tot (mTI + mTE), respiratory rate [RR = 1/(mTI + mTE)], and total ventilation [minute ventilation (E) = VT · RR].

Intrinsic positive end-expiratory pressure (PEEPi) was measured as the amplitude of the negative deflection of Pes between the onset of inspiratory effort and the onset of inspiratory flow. Respiratory motor output was evaluated on the basis of the transdiaphragmatic pressure-time product (PTPdi).

We computed PTPdi as the area subtended by transdiaphragmatic pressure (Pdi; Pdi = gastric pressure  Pes) above the end-expiratory baseline over inspiratory time (16).

PTPdi was multiplied by the respiratory frequency and expressed in centimeters of water per second per minute.

In addition, the positive deflection of Pdi (Pdi) was measured at 250 (Pdi₂₅₀) and 500 ms (Pdi₅₀₀) after the onset of inspiratory activity, detected by the onset of esophageal negative deflection.

Central respiratory output was evaluated from EMGdi. The EMGdi signal was band-pass filtered between 20 Hz and 1 kHz. The artifacts caused by the electrocardiogram were manually gated off. In addition, the signal was rectified and was processed in two different ways according to Lopata et al. (19), involving 1) integration to obtain the total electrical activity per breath (integrated EMGdi) and 2) averaging over 200-ms intervals to obtain a moving-time average signal.

Because the EMGdi signal continued into expiration during control periods and because end inspiration may be passive during IPSV (2), integrated activity per breath was determined not only during mTI but also during T_tot. These parameters are referred to as inspiratory-integrated EMG (EMGdi,mTI) and total integrated EMGdi (EMGdi,T_tot). In addition, because respiratory frequency might change across periods and because we wished to quantify diaphragm electrical activity independently from T_tot, integrated EMG activity was also expressed per minute (EMGdi,min). Furthermore, the peak inspiratory amplitude of the moving-time average signal (EMGdi_peak) and the electrical inspiratory time (eTI), as determined from onset to peak EMGdi, were also measured.

EMGdi parameters were expressed as percentages of the mean of the two control periods.

Individual mean values were calculated for each variable in a given session by averaging the breath-by-breath variables during the last 2 min of the recordings for each period.

Possible changes in end-expiratory volume associated with IPSV were analyzed by using inductive plethysmography when control 1 was abruptly replaced by an IPSV period (IPS or IPS+CO₂) and when an IPSV period was abruptly replaced by control 2.

Statistical Analysis

Concerning the EMG data, statistical differences among the four periods were tested by using the nonparametric Friedman test. The 5% level was chosen as significant. When a significant difference was observed, bilateral comparisons were performed by using the Wilcoxon signed-rank test.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Figure 1 shows examples of recordings during control 1, IPS, and IPS+CO₂ periods.

View larger version (55K): [in this window] [in a new window]   Fig. 1.   Tracing of end-tidal PCO2 (PETCO2), diaphragmatic EMG activity (EMGdi), flow, airway pressure (Paw), and transdiaphragmatic (Pdi) and esophageal pressures (Pes) during control, inspiratory pressure support (IPS), and IPS+CO2 in a representative subject. Note that, compared with control situation, EMG activity and swings in Pdi and Pes decreased during IPS and remained lower than control values during IPS+CO2.

Analysis of variance for repeated measurements did not detect any order effect.

No changes in end-expiratory volume were observed when control 1 was abruptly replaced by an IPSV period (IPS or IPS+CO₂), or when an IPSV period was abruptly replaced by control 2. In addition, PEEPi was consistently <1 cmH₂O, suggesting that all subjects were near the functional residual capacity in all the conditions.

The ventilatory pattern, and PET_CO2, PTPdi, Pdi₂₅₀, and Pdi₅₀₀ values observed in each condition are given in Table 1.

PTPdi individual data are shown in Fig. 2.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Individual values of transdiaphragmatic pressure-time product (PTPdi) during control 1, IPS, IPS+CO2, and control 2 periods. No significant differences were observed between control 1 and control 2, whereas PTPdi of IPS and IPS+CO2 were significantly lower than PTPdi of control 1 and control 2.

No differences were found between the two control periods.

IPS induced a significant increase in VT with no effect on RR (Table 1). Consequently, IPS induced a significant increase in E and a significant decrease in PET_CO2 (Table 1).

Decreases in PTPdi, Pdi₂₅₀, and Pdi₅₀₀ were also seen during IPS (Fig. 2, Table 1).

During IPSV, returning PET_CO2 to normal (IPS+CO₂) was associated with small but significant increases in the mTI-to-T_tot ratio, a slight, nonsignificant increase in RR, and a significant increase in E (with no effect on VT) (Table 1). The small E increase was associated with small, insignificant increases in PTPdi, Pdi₂₅₀, and Pdi₅₀₀ (Table 1, Fig. 2). However, these parameters remained lower during IPS+CO₂ than during the control periods, although PET_CO2 values were identical (Table 1, Fig. 2).

The main results regarding EMGdi are shown in Fig. 3. We found no differences between IPS and IPS+CO₂. In contrast, all expressions of the integral of EMGdi and eTI during IPS+CO₂ remained lower than during the control periods, despite identical PET_CO2 values (Fig. 3). EMGdi_peak was reduced during IPS and during IPS+CO₂ compared with the control periods (73 ± 8% and 65 ± 6%, respectively), but these differences did not reach statistical significance.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Individual values of EMGdi expressed as %control values for control 1, IPS, IPS+CO2, and control 2. EMGdi was quantified by integrating EMGdi during mechanical inspiratory time (EMGdi,mTI), total respiratory time (EMGdi,Ttot), and per minute and also by calculating electrical inspiratory time (eTI) of moving-time average of EMG signal. No significant differences were observed between control 1 and control 2, whereas IPS and IPS+CO2 values were significantly lower than control 1 and control 2 values for each parameter, except EMGdi,Ttot, which was not different between IPS and either control situation.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Breathing is a complex motor activity originating in a central neural drive located in the medulla and pons. This central neural drive is controlled via reflexes involving specialized organs that sense body fluid composition changes reflecting O₂ demands and CO₂ production. The sensory organs involved in these homeostatic reflexes are peripheral arterial chemoreceptors and central chemoreceptors. However, reflexes initiated by chemoreceptors, i.e., the chemical control of breathing, are not the only mechanisms responsible for the control of breathing. Respiratory control is also influenced by other regions of the brain and by peripheral afferent information from receptors in the airways and respiratory muscles. The resulting pattern of control is efficient and responsive to change. It has been suggested that the peripheral nonchemical receptors sense changes in mechanical conditions of ventilatory loading and adjust the breathing pattern to economize work of breathing (7). In humans, studies of ventilatory loading have found evidence of nonchemical control of respiratory activity (7). A nonchemical influence on inspiratory activity has also been observed during ventilatory unloading by volume-controlled mechanical ventilation (1). As a result of this nonchemical influence, the thoracic displacement induced by volume-controlled mechanical ventilation (CMV) may exert an inhibitory influence on inspiratory activity (1).

Over the last few years, partial ventilatory support techniques such as IPSV and proportional assist ventilation (PAV) have generated considerable interest as an alternative to volume CMV. Partial ventilatory support techniques decrease inspiratory activity, although it is the patient's inspiratory effort that triggers the ventilator-assisted breath. However, the mechanisms by which partial ventilatory assistance reduces inspiratory neuromuscular output have recently been questioned, and evidence has been found suggesting that nonchemical inhibition of inspiratory activity may exist during IPSV (26).

Shams and Scheid (26) found that anesthetized cats receiving IPSV exhibited a decrease in respiratory drive as assessed by EMGdi and demonstrated that this decrease was due not only to the induction of hypocapnia but also to influences from vagal afferents. When hypocapnia was corrected by CO₂ inhalation, the amplitude of the iEMG remained reduced during IPSV in comparison with normal breathing. However, these results may not be generalizable to humans because vagal influences in humans are fairly weak. In a study of intensive-care-unit patients, Bonmarchand et al. (4) found that improvement in the shape of the pressure increase during IPSV was associated with a decrease in diaphragmatic activity, although Pa_CO2 remained unchanged. This finding suggests that nonchemical inhibition of inspiratory activity exists during IPSV and is influenced by the shape of the pressure increase. However, this study was performed in patients with acute respiratory failure, and its results may not apply to patients with chronic respiratory failure or to healthy subjects. Scheid et al. (25) evaluated respiratory responses to inhaled CO₂ in normal human subjects during IPSV and found that IPSV effectively increased ventilation and reduced inspiratory activity at any given PET_CO2. However, they assessed respiratory center output on the basis of the occlusion pressure P_0.1, which reflects the initial part of the respiratory drive and is only minimally informative as to what happens later during inspiration. Morrell et al. (20) observed that during non-rapid-eye-movement sleep IPSV caused a decrease in EMGdi, in the absence of any changes in VT, PET_CO2, or respiratory frequency. However, they found no consistent effects on EMGdi in awake subjects. The discrepancies between the findings of Morrell et al. and Scheid et al. may be ascribable to differences in the IPSV devices used. For example, we recently demonstrated (16) that the device used by Morrell et al. was less efficient in reducing inspiratory activity than the one used by Scheid et al.

To further explore in humans and during wakefulness the existence of nonchemical inhibition of ventilatory activity, we conducted a study in healthy volunteers receiving IPSV, in whom 1) the PET_CO2 level was controlled by adding CO₂ to the inspired circuit and 2) both the work performed by the diaphragm and the electrical activity of the diaphragm were evaluated.

The reason for the reduction in PTPdi, Pdi₂₅₀, and Pdi₅₀₀ during IPS and IPS+CO₂ compared with the control conditions is unclear because these indexes reflect not only neural drive but also a number of other factors, including velocity reflected by VT-to-mTI ratio and lung volume changes modified by IPSV. Therefore, we also recorded EMGdi, which provides a more direct assessment of central respiratory output. Because EMGdi is affected by changes in lung volume (8), the recording of EMGdi may not be representative of the central respiratory output. We did not observe changes in end-expiratory volume when control was abruptly replaced by an IPSV period and when the IPSV period was replaced by control. However, an absence of any change in end-expiratory lung volume during transition does not necessarily mean that this parameter remained constant throughout. We therefore studied PEEPi, which remained under 1 cmH₂O during all trials, suggesting that end-expiratory lung volume remained in the vicinity of the functional residual capacity in all conditions. In addition, Gandevia and McKenzie (8) demonstrated that inspiratory output, represented by EMGdi, is overestimated when lung volume is increased, whereas we observed a decrease in EMGdi despite an increase in VT during IPSV. We can therefore assume that the decrease in EMGdi during IPSV cannot be explained by a change in lung volume. Moreover, before starting this study, we tested various conditions of diaphragmatic activity recording, including anchoring of the EMG catheter with an inflated gastric balloon as a means of stabilizing the physical relationship between the electrodes and diaphragm (unpublished observations). We drew conclusions similar to those of Önal et al. (21), who found that changes in electrode position had minimal effects on EMG quantification and that stabilization of the catheter did not improve the reproducibility of EMG data. On the basis of these considerations, we believe that EMG data obtained under the conditions used in our study provide reasonably reliable data on the central neuromuscular output to the diaphragm.

Another key issue is the accuracy of PET_CO2 to track arterial CO₂, especially during mechanical ventilation. Unfortunately, we did not measure the changes in arterial-end-tidal CO₂ difference in the different experimental conditions. However, Simon et al. (29) have previously demonstrated no change in arterial-end-tidal CO₂ difference when spontaneous eupnea was compared with a situation of mechanical hyperventilation (with VT similar to those observed in our study) and isocapnic condition induced, as in our study, by adding CO₂ to the inspiratory circuit. Therefore, we believe that PET_CO2 yields a valid estimate of arterial CO₂ in our study.

Although our subjects had neither previous knowledge of the issue under investigation nor previous experience with respiratory experimentation, we cannot exclude a voluntary response effect. For example, as in other studies (2, 12, 17, 25), the absence of respiratory rate reduction during IPS in the face of major hypocapnia may be ascribable to a voluntary response. However, when we corrected the hypocapnia by adding CO₂ to the inspiratory circuit, we found that the level of respiratory activity still remained lower than during the control periods. This result is in keeping with a study by Morrell et al. (20), in which IPS during non-rapid-eye-movement sleep caused a decrease in EMGdi in the absence of changes in VT, PET_CO2, or respiratory frequency, suggesting that, during IPSV, nonchemical influences on breathing may be initiated during sleep, i.e., in the absence of voluntary responses. Our study suggests that these nonchemical influences may persist during wakefulness, bearing out the hypothesis put forward by Scheid et al. (25) and supporting earlier findings in animals (26) and human patients (4).

The IPS device chosen was one of the best devices presently used in intensive care units (5). Nevertheless, we observed some dyssynchrony between the onset of inspiratory activity and the onset of IPSV (see Fig. 1). This may influence the response to IPSV via an increase in inspiratory activity during IPSV, which is responsible for a decrease in the nonchemical influence of mechanical ventilation. Despite this potential "instrumental artifact," we found evidence of nonchemical inhibition of inspiratory activity during IPSV, and the expected physiological effect was not obscured.

Air and the CO₂-corrected mixture were tested in random order. Conceivably, when IPS+CO₂ follows a trial of IPS, the hypocapnia during the previous IPS session may continue to influence respiratory motor output for some time during the IPS+CO₂ session. However, our analysis compared minutes 10 and 12 of each period, and we consistently found that respiratory parameters stabilized before minute 8 of each period. In addition, our statistical evaluation takes into account the order in which subjects received IPS and IPS+CO₂. This order failed to significantly influence the results.

When PET_CO2 was kept normal during IPSV via CO₂ inhalation, RR increased nonsignificantly, E increased slightly but significantly, VT remained unchanged, and respiratory motor output indexes showed nonsignificant increases. The increase in E was ~0.2 l · min¹ · mmHg¹, which is a relatively weak response. Some previous studies found no significant change in E until PET_CO2 reached hypercapnic levels (12, 25), whereas Georgopoulos et al. (9) observed, as in our study, a weak ventilatory response (of ~0.35 l · min¹ · mmHg¹) in the hypocapnic range and a much stronger ventilatory response in the hypercapnic range (see Fig. 4 in Ref. 9).

Nonchemical inhibition of inspiratory activity was not observed in a study of PAV performed by Georgopoulos et al. (10). They demonstrated that during PAV the decrease in respiratory motor output was due only to the change in PET_CO2. Direct comparison of our study to that by Georgopoulos et al. is difficult. First, the study by Georgopoulos et al. was performed during a dynamic condition of ventilatory CO₂ response, whereas our condition was static. Second, Georgopoulos et al. evaluated their subjects in a hypercapnic range, from 50 to 59 Torr, whereas we evaluated IPSV in hypocapnic and normocapnic conditions (differences in CO₂ responses between the hypocapnic and hypercapnic ranges have been discussed above). Third, PAV and IPSV are obviously not similar modes of ventilation. For example, after the initial active triggering phase during IPSV, inspiration can be virtually passive and VT can be above physiological values despite minimal inspiratory activity (2), whereas during PAV the assistance is proportional to respiratory muscle activity and VT depends on the duration and the strength of respiratory muscle activity. These differences in experimental techniques and ventilatory modes may account for the apparent discrepancy between our findings and those of Georgopoulos et al.

Nonchemical inhibition of inspiratory activity has been observed during CMV. Puddy et al. (23) observed a small, nonchemical inhibition of inspiratory activity during CMV. However, their conclusion was based only on an analysis of the respiratory rhythm. Henke et al. (11) reported that returning PET_CO2 to the spontaneous level by adding CO₂ to the inspired air during mechanical hyperventilation did not completely restore respiratory activity to the level observed during spontaneous breathing. Altose et al. (1), Simon et al. (27, 29), and others (13, 18, 22) found that, after suppression of inspiratory activity by mechanical hyperventilation, the level of PET_CO2 at which inspiratory activity was detectable was several Torr above the eupneic PET_CO2. This suggests a nonchemical inhibitory effect of CMV similar to that seen during IPSV. Some of these studies were performed in sleeping subjects (11, 13, 27), indicating that these nonchemical inhibitory effects are largely of peripheral origin and do not depend on higher brain centers.

In keeping with these studies, we suggest that the thoracic displacement imposed by IPSV exerts a nonchemical inhibitory influence, the origin of which is probably in peripheral afferents. In our study, the reduction in inspiratory activity during IPSV was due primarily to a reduction in the duration of inspiratory activity as assessed on the basis of eTI (see Fig. 3; compared with control periods, IPS+CO₂ was associated with a decrease in eTI and with a nonsignificant decrease in the peak of electrical activity). Reduction in inspiratory activity may also be due to the reduction in the slope of this inspiratory activity, according to the significant decrease in Pdi₂₅₀ and Pdi₅₀₀ and to the nonsignificant decrease in the EMGdi_peak. This leads us to believe that the lung inflation and VT-to-mTI ratio increase induced by IPSV during neural inspiration reduce and/or terminate neural inspiration.

Among studies done to investigate the mechanism of nonchemical inhibition of mechanical ventilation in humans, some compared the response to mechanical ventilation between normal subjects and patients with loss of a sensory pathway (15, 18, 28, 29). The nonchemical inhibition associated with mechanical ventilation persisted in C₄-C₅ quadriplegics with intercostal deafferentation (28). Studies in lung-transplant patients with bilateral vagal denervation yielded conflicting results: persistence of nonchemical inhibition during mechanical ventilation was observed by Simon et al. (29) and by Leevers et al. (15) but not by Lofaso et al. (18). Thus the pathways of nonchemical inhibition associated with mechanical ventilation remain unclear.

In conclusion, our data support the existence of nonchemical inhibition of inspiratory activity during IPSV. This may explain why IPSV is effective in patients with imminent failure of respiratory muscles in the absence of any concomitant improvement in arterial blood-gas values.