HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wilson, C. R. Articles by Dempsey, J. A. Search for Related Content PubMed PubMed Citation Articles by Wilson, C. R. Articles by Dempsey, J. A.

Non-chemical inhibition of respiratory motor output during mechanical ventilation in sleeping humans

Christine R. Wilson *, Makoto Satoh *, James B. Skatrud ¹ and Jerome A. Dempsey *

MS 9007 Received 26 November 1998; accepted after revision 9 April 1999.

	ABSTRACT

Top Abstract Introduction Methods Results Discussion References

1. To determine the magnitude and time course of changes in respiratory motor output caused by non-chemical influences, six sleeping subjects underwent assist-control mechanical ventilation (ACMV) at increased tidal volume (V_T). During ACMV, end-tidal P_CO2 (P_ET,CO2) was either held at normocapnic levels (P_ET,CO2, 0·6-1·1 mmHg > control) by adding CO₂ to the inspirate, or it was allowed to fall (hypocapnia).

2. Each sleeping subject underwent several repeat trials of twenty-five ACMV breaths (V_T, 1·3 or 2·1 times control; peak flow rate, 30-40 l min^-1; inspiratory time, ± 0·3 s of control). The end-tidal to arterial P_CO2 difference throughout normocapnic ACMV at raised V_T was unchanged from eupnoeic levels during studies in wakefulness.

3. Normocapnic ACMV at both the smaller and larger increases in V_T decreased the amplitude of respiratory motor output, as judged by decreased maximum rate of rise of mask pressure (P_m) (mean dP_m/dt_max, 46-68 % of control), reduced diaphragmatic EMG (to 55 % of control) and reduced V_T on the first spontaneous breath after ACMV (to 70 % of control). Expiratory time (T_E) was slightly prolonged (13-32 % > control). This inhibition of amplitude of respiratory motor output progressed over the first five to seven ventilator cycles, was maintained over the remaining 18-20 cycles and persisted for three to five spontaneous breaths immediately following cessation of ACMV.

4. Hypocapnia did not further inhibit respiratory motor output amplitude beyond the effect of normocapnic ACMV at high V_T, but did cause highly variable prolongation of T_E when P_ET,CO2 was reduced by greater than 3 mmHg for at least five ventilator cycles.

5. These data in sleeping humans support the existence of a significant, non-chemical inhibitory influence of ACMV at increased V_T and positive pressure upon the amplitude of respiratory motor output; this effect is manifested both during and following normocapnic mechanical ventilation.

	INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

Is non-chemical or mechanoreceptor feedback important in regulating the amplitude and timing of respiratory motor output in humans? Classically, breath holding experiments were used by Fowler (1954) to demonstrate that higher levels of CO₂ could be tolerated in humans while breathing than while breath holding. More recently, non-chemical regulation of breathing in humans has been addressed by using mechanical ventilation to alter tidal volume (V_T), breathing frequency, flow rate and/or levels of pressure support. During wakefulness, the reported effects of mechanical ventilation and increasing V_T have been highly variable, ranging from major reductions in breathing frequency, to negligible effects on breath timing (Simon et al. 1992; Leevers et al. 1994; Manchanda et al. 1996; Puddy et al. 1996; Georgopoulos et al. 1996, 1997; Tobert et al. 1997). However, wakefulness probably heightens one's awareness of sensory inputs (inputs which are substantial during mechanical ventilation), which may well over-ride even marked inhibitory influences on respiratory motor output, including hypocapnia (Fink et al. 1963; Skatrud & Dempsey, 1983; Datta et al. 1991; Corfield et al. 1995).

Fewer studies have been conducted during non-rapid eye movement (NREM) sleep and these are also inconsistent. Some findings show a significant effect of mechanical ventilation and/or increased V_T on respiratory motor output (Henke et al. 1988; Morrell et al. 1993; Leevers et al. 1994; Scheid et al. 1994; Manchanda et al. 1996), while others claim no effect of mechanical ventilation, per se, on the timing and/or amplitude of respiratory motor output which was independent of reduced P_CO2 (Skatrud & Dempsey, 1983; Henke et al. 1988; Tobert et al. 1997; Meza et al. 1998). These studies, which used mechanical ventilation during sleep, have not considered variations in V_T amplitude or the effects of duration of mechanical ventilation, distinguished between the effects of mechanical ventilation on amplitude versus timing of respiratory motor output or contrasted the effects of raised V_T, per se, from hypocapnia.

We proposed a comprehensive approach to this question of non-chemical inhibition by: (a) determining the effects of graded increases in V_T on respiratory motor output during NREM slow-wave sleep while normocapnia was maintained; (b) using the assist-control mode of mechanical ventilation (ACMV) in order to assess the time course and after-effects of raised V_T on changes in both timing and amplitude of respiratory motor output; (c) comparing normocapnic and hypocapnic conditions to determine any interactive effects of mechanical and chemical influences on respiratory motor output. We employed three methods for assessing respiratory motor output and used serial arterial blood sampling to determine the accuracy with which systemic normocapnia was maintained during ACMV.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Study 1. To determine the time course of inhibition of inspiratory motor output during mechanical ventilation in sleeping humans

Subjects. Six healthy subjects (2 men and 4 women; 19-23 years old) who, by self report, were heavy sleepers, non-snorers and with no history of respiratory or cardiovascular disease were recruited from the university community and studied during NREM sleep. These experiments were performed in accordance with the Declaration of Helsinki, approved by the Human Studies Committee at the University of Wisconsin Centre for Health Sciences, Madison, and informed consent was obtained from all subjects.

Instrumentation. Subjects were studied at night following sleep deprivation ( 2 h of sleep) on the previous night. Each subject was instrumented with electrodes to record electroencephalographic activity (EEG: C3/A2, C4/A1, O1/A2, O2/A1), chin electromyographic activity and eye electro-oculographic activity. In addition, surface electrodes were placed on the chest to record electrocardiographic activity (ECG, modified lead II) and diaphragm EMG activity (EMG_di; electrodes were placed 2-4 cm above the costal margins, between the mid-clavicular and mid-axillary lines). Respiratory inductance plethysmograph (Respitrace, Ambulatory Monitoring, Inc., Ardsley, NY, USA) elastic bands were placed around the chest and abdomen and recordings were made in the DC mode.

An appropriately sized nasal CPAP mask (Respironics, Inc., Pittsburgh, PA, USA) was fitted to each subject. Attached to the mask was a heated pneumotachograph (model 5719, 0-100 l min^-1; Hans Rudolph, Inc., Kansas City, MO, USA) connected to a differential pressure transducer (± 0·8 cmH₂O; Validyne Engineering Corp., Northridge, CA, USA) to measure bidirectional airflow. The inspiratory and expiratory lines of the ventilator tubing were connected to the pneumotachograph using a Y connector with one-way valves attached to either side (1-2 cmH₂O opening pressure). Mask pressure (P_m) was measured through a port in the mask (± 56 cmH₂O; Validyne), and a second mask port was used to monitor P_ET,CO2 (CD-3A; AEI Technologies, Inc., Pittsburgh, PA, USA).

Leaks were identified by comparing inhaled and exhaled tidal volumes and by using a CO₂ sensor around the perimeter of the mask. In addition, the V_T indicated by the summed Respitrace signal was compared with the V_T indicated by electronic integration of the pneumotachograph flow signal in order to confirm the absence of leak and also to estimate any changes in end-expiratory lung volume. Theatrical glue and putty were used to seal mask leaks and subjects' mouths were taped to prevent air leakage through the mouth.

We quantified inspiratory motor output by measuring the maximal rate of change in mask pressure (dP_m/dt_max) at the beginning of each inspiratory effort (Matthews & Howell, 1975). All measurements were taken at < 150 ms and were isovolumetric (i.e. measured before the inspiratory valve opened). The diaphragmatic EMG (EMG_di) signal was rectified, filtered and integrated (100 ms time constant), and the mean EMG_di area was measured by digitizing (Zidas, Carl Zeiss, Inc., Thornwood, NY, USA) that portion of the signal which did not contain ECG artefacts and dividing the total area by the duration of the signal.

Protocol. Trials were performed during stages III and IV NREM sleep. A trial consisted of 10 spontaneous breaths (ventilator mode: CPAP = 0 cmH₂O), 25 ventilator breaths (assist-control mode), and five recovery spontaneous breaths (CPAP = 0 cmH₂O). When the mechanical ventilator (Adult Star, Mallinckrodt, Inc., Carlsbad, CA, USA) was set in the assist-control mode (ACMV), the trigger sensitivity was 1-2 cmH₂O and the ventilator frequency was 0·5 breaths min^-1. The ventilator V_T was set at either 10-30 % (low V_T trial) or 80-125 % (high V_T trial) above spontaneous V_T and the peak flow rate was set at 30-40 l min^-1; these settings resulted in a decrease in inspiratory time (T_I, compared with baseline values) during the low V_T trials, and a prolongation of T_I during the high V_T trials. End-tidal CO₂ was either maintained at spontaneous breathing levels by adding CO₂ to the inspirate (normocapnic trial) or allowed to fall (hypocapnic trial). The ventilator was switched during expiration to the spontaneous breathing mode after either 25 respiratory cycles or at the onset of periodic breathing.

If a subject had difficulty falling asleep or consistently awakened during ventilator trials, a hypnotic was given orally (0·125 mg triazolam, supplemented by an additional dose if necessary). Forty-two trials were obtained in five subjects without triazolam and 29 trials were obtained in five subjects where triazolam was used. Changes in dP_m/dt_max and breath timing during ACMV were not different within a subject during NREM sleep with and without triazolam. Therefore, findings from all sleep trials were combined.

Data analysis. Data were recorded on videotape for subsequent computer acquisition and analysis using software programs developed in our laboratory. Inspiratory and expiratory V_T were determined by digital integration of the flow signal. Respiratory frequency was determined from the mask pressure signal; weak inspiratory efforts were identified by noting a small negative deflection in the mask pressure and/or increased EMG_di activity.

The EEG tracings were carefully reviewed to determine sleep stage. Trials were only included in the analysis if (i) the subject was in stage III or IV sleep throughout the trial, and (ii) there was no evidence of arousal (EEG alpha activity > 3 s in duration).

For each subject, all trials performed under a given condition (e.g. normocapnia, low V_T) were averaged to obtain a representative value for that subject for each measured variable. The representative group mean values were then derived based on one value per subject.

Statistical analysis. Group mean measurements made during spontaneous breathing and mechanical ventilation were compared using two-way repeated-measures analysis of variance and Dunnett's post hoc test (Tables 1 and 2). The after-effect of ACMV on respiratory motor output was determined by comparing the first recovery breath with the baseline value using Student's paired t test. Correlations between the three different indices of respiratory motor output during and following ACMV were determined by linear regression.

Study 2. To determine the difference between end-tidal and arterial P_CO2 during normocapnic mechanical ventilation and spontaneous breathing

Subjects and instrumentation. Five healthy subjects (4 men and 1 woman, 18-32 years) were studied while awake. Subjects were instrumented as in Study 1. Each subject also had an arterial catheter placed in either a radial or brachial artery following local lidocaine (lignocaine) anaesthesia. Arterial blood parameters were analysed (Radiometer Corp) using tonometered blood for calibration.

Protocol. Subjects underwent five trials of ACMV; the ventilator frequency was set equal to the subject's eupnoeic frequency and the ventilator V_T was set so that the inspired CO₂ fraction (F_I,CO2) during normocapnic trials matched that during the sleep studies in Study 1 (F_I,CO2 = 0·009 ± 0·0002). In each subject, 2 ml arterial blood samples were repeatedly taken during four normocapnic ventilator trials at raised V_T and one hypocapnic ventilator trial (the last trial). Each trial consisted of 2 min of spontaneous breathing followed by 2 min of ACMV. Four blood samples were taken during spontaneous breathing and four samples during ACMV; each sample was drawn over 30 s. During mechanical ventilation, blood sampling began after the third breath delivered by the ventilator. End-tidal P_CO2 was measured for the breaths occurring during each arterial sample, and then averaged to obtain the P_ET,CO2 per sample.

Statistical analysis. One-way repeated-measures analysis of variance was used to compare the arterial - end-tidal P_CO2 differences during spontaneous breathing and mechanical ventilation.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Time course of changes with raised V_T in normocapnia during assist-control mechanical ventilation (ACMV)

A typical breath-by-breath response to twofold increases in V_T above spontaneous breathing during NREM sleep in maintained normocapnia is shown in Fig. 1. Note the slight slowing of breathing frequency and reductions in the amplitude of both EMG_di and dP_m/dt_max in the first few ventilator cycles, and further reductions in the EMG_di and dP_m/dt_max as normocapnic mechanical ventilation at raised V_T was continued for twenty-five cycles. Then, following abrupt cessation of mechanical ventilation, the first recovery breath showed a markedly reduced V_T, EMG_di and dP_m/dt_max followed by their slow recovery to control levels.

		View larger version [in this window] [in a new window]
Figure 1. Recording of one trial of normocapnic ACMV at high ventilator volume (850 ml, 215 % baseline VT) during stage III sleep Note: (a) the diminution of EMGdi activity and dPm/dtmax and slight slowing of respiratory rate during mechanical ventilation and (b) the slow recovery of spontaneous VT after mechanical ventilation was discontinued. Dashed vertical lines indicate the start and finish of mechanical ventilation.

Group mean time course responses are shown (in Fig. 2 and Table 1) for all trials of normocapnic ACMV during which V_T was increased an average of either 1·28 (Fig. 2A) or 2·10 (Fig. 2B) times the mean values obtained during spontaneous breathing and mean P_ET,CO2 was held 0·6-1·1 mmHg above control (P < 0·05). In order to increase V_T, V_T/T_I was increased to 1·6 and 1·8 times the eupnoeic value at low and high V_T, respectively; T_I was reduced to 0·3 s less than eupnoeic T_I at the low V_T and increased to 0·3 s greater than eupnoeic T_I at the high V_T. End-expiratory lung volumes were unchanged from control during ACMV.

		View larger version [in this window] [in a new window]
Figure 2. Group mean data during normocapnic ACMV VT, peak VI (inspiratory flow); Ttot, respiratory cycle time; dPm/dtmax, maximum rate of change of mask pressure; PET,CO2, mean area EMGdi (diaphragmatic EMG). Data represent means ± S.D. for six subjects, one to six trials per subject. A, normocapnic ACMV at low ventilator VT (128 % baseline VT) (n = 6). As one subject was aroused from sleep, only five subjects are represented in data beyond ventilator breath 22 (note break in horizontal lines). Mean area EMGdi data represent the means ± S.D. of one trial in each of the six subjects. B, normocapnic ACMV at high ventilator VT (210 % baseline VT) (n = 6). Mean area EMGdi data represent the means ± S.D. of one trial in each of five subjects (1 trial per subject). Only four subjects are represented in data beyond ventilator breath 23. Dashed vertical lines indicate start and finish of mechanical ventilation.

Table 1. Effects of ACMV at increased V_T in sustained normocapnia on respiratory motor output

	VT (ml)	VT/TI (l s-1)	TI (s)	TE (s)	Ttot (s)	dPm/dtmax (cmH2O s-1)	Mean area EMGdi (%)	PET,CO2 (mmHg)
A. Low VT
Baseline	417	0·29	1·5	2·2	3·6	13·2	100	40·8
	(7·7)	(0·007)	(0·03)	(0·05)	(0·06)	(0·81)	-	(0·11)
Mechanical ventilation
Breath 5	536 **	0·47 **	1·2 **	2·4 **	3·5	11·4 **	88 **	41·4 **
	(27·6)	(0·073)	(0·18)	(0·25)	(0·32)	(2·01)	(24·6)	(2·54)
Breath 10	535 **	0·45 **	1·2 **	2·5 **	3·7	9·0 **	69 **	41·9 **
	(27·6)	(0·065)	(0·18)	(0·32)	(0·46)	(2·12)	(15·7)	(2·83)
Breath 15	531 **	0·45 **	1·2 **	2·5 **	3·7	9·2 **	71 **	41·2 **
	(23·7)	(0·062)	(0·16)	(0·30)	(0·41)	(1·80)	(22·4)	(2·49)
Breath 20	532 **	0·46 **	1·2 **	2·5 **	3·7	9·0 **	56 **	41·4 **
	(25·7)	(0·059)	(0·14)	(0·26)	(0·37)	(3·37)	(7·9)	(2·64)
B. High VT
Baseline	420	0·28	1·5	2·2	3·7	14·2	100	40·6
	(5·8)	(0·003)	(0·01)	(0·03)	(0·03)	(0·44)	-	(0·13)
Mechanical ventilation
Breath 5	895 **	0·50 **	1·8 *	2·7 **	4·5 **	8·7 **	72 **	41·5 **
	(4·4)	(0·076)	(0·27)	(0·42)	(0·65)	(1·70)	(7·1)	(3·06)
Breath 10	888 **	0·48 **	1·9 **	2·8 **	4·7 **	6·8 **	54 **	41·4 **
	(4·9)	(0·083)	(0·36)	(0·42)	(0·73)	(2·33)	(15·0)	(2·98)
Breath 15	885 **	0·48 **	1·9 **	2·8 **	4·7 **	7·0 **	65 **	41·3 **
	(5·7)	(0·073)	(0·27)	(0·50)	(0·72)	(2·35)	(20·2)	(2·71)
Breath 20	875 **	0·47 **	1·9 **	2·8 **	4·7 **	6·7 **	54 **	41·5 **
	(6·4)	(0·085)	(0·36)	(0·30)	(0·61)	(1·96)	(10·6)	(2·85)
Breath 25	873 **	0·47	2·0 **	2·9	4·9 **	6·6	-	41·8
	(5·8)	(0·093)	(0·45)	(0·72)	(1·00)	(2·67)	-	(3·32)
Recovery breath 1	292 **	0·22	1·3	2·1	3·5 **	7·8	-	-
	(8·8)	(0·062)	(0·14)	(0·46)	(0·57)	(2·67)	-	-

Respiratory cycle time (T_tot) increased to 15 ± 6 % above control T_tot (mean ± S.D., P < 0·01 compared with baseline) immediately upon the first ventilator cycle at high V_T, and rose slightly further with time to 26 ± 8 % above control at the tenth cycle, with no further changes over fifteen additional cycles of ACMV. T_tot fell to slightly less than control immediately on the first spontaneous breath following cessation of ACMV.

The EMG_di and dP_m/dt_max were not measurably affected on the first cycle of ACMV at the higher V_T but, beginning with the second cycle, showed a progressive decline which, by ventilator cycle 5, was reduced to 62 ± 9 % of control for dP_m/dt_max and 72 ± 7 % of control for EMG_di (P < 0·01, compared with baseline). By ventilator cycle 10, dP_m/dt_max and EMG_di were reduced to 49 ± 16 and 54 ± 15 %, respectively, of control. Although ACMV at the higher V_T appeared to cause a greater decrease in respiratory motor output (dP_m/dt_max and EMG_di) than ACMV at low V_T (Fig. 2A versus B), statistical testing showed no significant difference between the mean values. This lack of significant difference may have been due to the variability within this small sample. For example, during ventilator breath 20, ACMV at low V_T caused 40 % decrease in dP_m/dt_max in three of the six subjects whereas, during ACMV at high V_T, dP_m/dt_max decreased by > 40 % in all subjects.

Group mean values for the first spontaneous recovery V_T following cessation of ACMV were 69 ± 20 % of the eupnoeic control V_T following cessation of high V_T ACMV and 72 ± 12 % following low V_T ACMV (P < 0·05). Both EMG_di and dP_m/dt_max were also reduced to about 40-50 % less than control on the first recovery breath (P < 0·05). V_T, EMG_di and dP_m/dt_max all gradually returned to control levels over four to six spontaneous breaths in recovery.

During ACMV at high normocapnic V_T the changes in both the magnitude and time course of inhibition of EMG_di and dP_m/dt_max were quite uniform among all six subjects and for repeat trials within the same subject. For the smaller increase in V_T in normocapnia, both the time course and magnitude of inhibition of respiratory motor output were much more variable between trials and among subjects, but all subjects did show measurable reductions in EMG_di and dP_m/dt_max during ACMV (see Fig. 6A below for individual subject responses to normocapnic mechanical ventilation at the nadir of their response).

Time course of changes with raised V_T plus hypocapnia during ACMV

A typical breath-by-breath response to a twofold increase in V_T above spontaneous breathing during ACMV is shown in Fig. 3. When hypocapnia was allowed to accompany the increase in V_T, the EMG_di and dP_m/dt_max fell quickly and, by the fifth mechanical ventilator cycle, T_E was prolonged substantially and an unstable breathing pattern ensued until arousal occurred. Note that following the sixth ventilator cycle an inspiratory effort is barely detectable on the volume trace, but this effort was not sufficient to produce the 1-2 cmH₂O pressure change required to trigger a ventilator breath.

		View larger version [in this window] [in a new window]
Figure 3. Recording of one trial of hypocapnic ACMV at high ventilator volume (870 ml, 203 % baseline VT) during stage III sleep Note: (a) the slowing of respiratory rate and (b) the rapid fall in EMGdi, leading to an inspiratory effort insufficient to trigger a ventilator cycle, and an irregular respiratory rhythm. The trial ends in arousal from sleep.

Individual subject values for hypocapnic trials are shown in Fig. 4A and 4B and group mean values are shown in Table 2 for all ventilator cycles up to the point where either apnoea or an untriggered breath occurred. Thereafter, alveolar gases were highly variable as was the breathing pattern (see Fig. 3). With hypocapnic mechanical ventilation at both levels of raised V_T, P_ET,CO2 fell gradually, reaching a plateau after about five mechanical ventilator cycles at -1·9 ± 1·0 mmHg below the eupnoeic level for the smallest increase in V_T and -5·9 ± 1·3 mmHg less than the eupnoeic level for the highest V_T.

		View larger version [in this window] [in a new window]
Figure 4. Individual subject responses (mean of 1-5 trials in each of 6 subjects) to hypocapnic ACMV A, hypocapnic ACMV at low ventilator VT (128 % of baseline VT). Most subjects show a gradual fall in dPm/dtmax and EMGdi but no apnoea. However, one subject became apnoeic following ventilator breath 8 after PET,CO2 fell by 4 mmHg. B, hypocapnic ACMV at high ventilator VT (212 % of baseline VT). Note that only half of the subjects experienced two- to fivefold prolongations in Ttot; weak inspiratory efforts ended the trials of the remaining three subjects.

Table 2. Effects of ACMV at increased V_T in hypocapnia on respiratory motor output

	VT (ml)	VT/TI (l s-1)	TI (s)	TE (s)	Ttot (s)	dPm/dtmax (cmH2O s-1)	Mean area EMGdi (%)	PET,CO2 (mmHg)
A. Low VT
Baseline	420	0·29	1·4	2·2	3·6	13·4	100	40·4
	(5·6)	(0·005)	(0·02)	(0·03)	(0·03)	(0·42)	-	(0·12)
Mechanical ventilation
Breath 5	536 **	0·46 **	1·2 **	2·5	3·7	10·9 **	75 *	38·5 **
	(29·1)	(0·065)	(0·19)	(0·41)	(0·57)	(1·19)	(17·9)	(1·15)
B. High VT
Baseline	404	0·29	1·4	2·2	3·6	13·7	100	39·2
	(3·2)	(0·003)	(0·02)	(0·02)	(0·04)	(0·53)	-	(0·05)
Mechanical ventilation
Breath 5	855 **	0·43 **	2·0 **	5·3	7·3	6·8 **	46 **	33·3 **
	(2·8)	(0·071)	(0·41)	(4·76)	(4·77)	(2·20)	(9·9)	(1·83)

The changes in breath timing during hypocapnic ACMV were highly variable both within and among subjects. At the higher V_T, three subjects had a three- to fivefold increase in T_tot (at ventilator cycles 5-9, 5-8 mmHg fall in P_ET,CO2), two subjects had a 70-90 % increase in T_tot (at ventilator cycles 6-7, 4-7 mmHg fall in P_ET,CO2), and one subject increased T_tot by 35 % (at ventilator cycle 5, 6 mmHg fall in P_ET,CO2).

Prior to the initiation of apnoea, EMG_di and dP_m/dt_max decreased in all subjects to about 50 % of eupnoea by ventilator cycle 5-8 during the high V_T ACMV (Fig. 4B). At the low V_T, EMG_di and dP_m/dt_max fell to about 25 % below eupnoeic control by cycle 8-10 and remained there throughout the remaining fifteen cycles (see Fig. 4A).

Raised V_T effects in normocapnia versus hypocapnia

The effects of increased ventilator V_T in normocapnia versus hypocapnia on the longest T_tot and the nadir dP_m/dt_max achieved during ACMV are summarized for all trials in Fig. 5. The following should be noted. (a) There was an absence of substantial prolongation of T_tot in normocapnia even at V_T values 2 times control. (b) Substantial prolongation of T_E and T_tot occurred only when P_ET,CO2 fell more than 3 mmHg below control. (c) The prolongation of T_E and T_tot varied more than twofold across trials at any given level of hypocapnia (-4 to -8 mmHg P_ET,CO2). These marked differences in T_tot prolongation were not found to be attributable to systematic differences among subjects or to variations in sleep state, duration of hypocapnia or magnitude of V_T during ACMV. (d) The magnitude of the reduction in nadir dP_m/dt_max during ACMV was not significantly affected by the amount of the fall in P_ET,CO2. (e) During small elevations in V_T, reductions in dP_m/dt_max were variable when normocapnia was maintained; however, dP_m/dt_max consistently fell when even mild hypocapnia accompanied the elevated V_T.

		View larger version [in this window] [in a new window]
Figure 5. Change in PET,CO2 versus longest Ttot and nadir dPm/dtmax Change in PET,CO2 from baseline levels for all normocapnic and hypocapnic trials is plotted against longest Ttot during each trial (A) and nadir dPm/dtmax occurring either at the same time as or before the breath with the longest Ttot (B). For the hypocapnic trials, the first significant prolongation of Ttot was used as the longest Ttot because following this initial Ttot prolongation PET,CO2 increases and both breathing pattern and PET,CO2 were unstable. , trials at low ventilator VT; , trials at high ventilator VT.

Some of the time course changes during ACMV were also affected by hypocapnia. With the larger increase in V_T, nadir values for EMG_di and dP_m/dt_max were reached more quickly during ACMV when hypocapnia accompanied the increase in V_T (Fig. 2B versus Fig. 4B, P < 0·05 at ventilator breath 5). However, for the smaller increase in V_T, the time course of reductions in EMG_di and dP_m/dt_max were similar in hypocapnia versus normocapnia (Fig. 2A versus Fig. 4A, P > 0·10 at ventilator breath 5).

Methods testing

As shown in Fig. 6A (individual trials) and Tables 1 and 2 (group mean values), reductions in EMG_di during ACMV correlated significantly and reasonably well with those in dP_m/dt_max. Both indices showed similar changes in group mean values (-50 % of control) and ranges of maximum inhibition (-20 to -90 % of control) and the relationship of changes in these parameters to each other was significant under each of the four conditions of ACMV.

		View larger version [in this window] [in a new window]
Figure 6. Assessment of indices of respiratory motor output A, nadir of the changes from baseline in mean area of EMGdi versus dPm/dtmax for 23 trials where EMG was analysed. The dotted line indicates the linear regression line (r = 0·61, P < 0·002). B, dPm/dtmax, VT and VT/TI for the first spontaneous breath after ACMV are shown for all trials. Dotted lines indicate linear regressions for dPm/dtmax versus VT and VT/TI (r = 0·73 (P < 0·001) and 0·65 (P < 0·001), respectively). Closed symbols, normocapnic ACMV; open symbols, hypocapnic ACMV; circles, low ventilator VT; squares, high ventilator VT.

We used spontaneous breaths which occurred immediately following ACMV without arousal to compare changes in EMG_di and dP_m/dt_max to changes in V_T. In the first recovery breath following almost all trials of ACMV all three of these indices of respiratory motor output were reduced below spontaneous breathing control values and all showed a similar slow return to control over five recovery breaths. However, the mean values for percentage reductions in V_T were usually less than those of either EMG_di or dP_m/dt_max. The correlation of reductions in V_T with those in dP_m/dt_max was significant (r = 0·73) as it was for V_T/T_I versus dP_m/dt_max (r = 0·65, Fig. 6B). The correlation of V_T (or V_T/T_I) with EMG_di on the first recovery breath was not significant although both variables were reduced below control following all but one ACMV trial.

The direction of change in dP_m/dt_max during and following ACMV and change in V_T following ACMV was reproducible both within and among subjects. For example, under normocapnic high V_T conditions, we conducted two to five ACMV trials in each of six subjects. Of the 16 total trials, all showed a greater than 30 % reduction in dP_m/dt_max during ACMV; spontaneous V_T after ACMV fell by > 10 % in 9 of the 11 trials where recovery breaths could be measured.

Figure 7 shows group mean values for arterial blood parameters obtained in five awake subjects during replication of the ACMV protocols carried out in sleep. The group mean arterial - end-tidal P_CO2 difference obtained under control conditions during spontaneous eupnoea remained unchanged (P > 0·10) at all sampling times throughout ACMV whether or not supplemental CO₂ was added to the inspirate. Of the total 80 measurements made of arterial - end-tidal P_CO2 difference during 'normocapnic' ACMV (4 measurements during each of 4 trials in 5 subjects), 61 changed by less than ± 1·0 mmHg from control, 13 changed by ± 1·1-1·5 mmHg from control and 6 changed by > 1·5 mmHg from control.

		View larger version [in this window] [in a new window]
Figure 7. PCO2, pH and the arterial - end-tidal PCO2 difference during eupnoea and ACMV Measurements of arterial PCO2 and pH and arterial - end-tidal PCO2 differences (means ± S.D.) in five awake subjects during normocapnic ACMV (4 trials per subject, ) and hypocapnic ACMV (1 trial per subject, ). Note that ACMV did not change the arterial PET,CO2 difference from baseline levels.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Summary of findings

Our study has addressed the question of non-chemical influences on the regulation of respiratory motor output in sleeping humans as represented by the time course of change in the timing and amplitude of EMG_di and in inspiratory dP_m/dt_max during assist-control mechanical ventilation at elevated V_T. Concerning our methods we showed: (a) that measurements of changes in EMG_di and dP_m/dt_max provided internally consistent estimates of changes in respiratory motor output amplitude which were also qualitatively similar to changes in spontaneous V_T following ACMV, and (b) that changes in end-tidal P_CO2 corresponded to changes in arterial P_CO2 during simulation of the ACMV protocol. ACMV at raised V_T in normocapnia and hypocapnia caused significant reductions in the amplitude of respiratory motor output which: (a) were not significantly affected by the magnitude of ventilator V_T; (b) fell progressively with time over the initial seven to ten ventilator cycles, reaching nadir values which were independent of the fall in P_ET,CO2; and (c) remained reduced over the initial few spontaneous breaths immediately following ACMV. Significant, but not substantial, T_E and T_tot prolongation occurred during large increases in V_T when normocapnia was maintained; however, T_E and T_tot were prolonged to a highly variable extent when P_CO2 was reduced greater than 3 mmHg for at least five ventilator cycles. We believe these data support the existence of a strong non-chemical inhibitory effect of ACMV at increased tidal volume on the amplitude of respiratory motor output in sleeping humans.

Measurement limitations

In order to test the effects of ACMV at raised V_T, we were dependent upon two indirect indices of respiratory motor output, both of which have several potential confounding factors. dP_m/dt_max reflects the rate of rise of inspiratory muscle activity when muscles contract isovolumetrically (Matthews & Howell, 1975). However, when interpreting changes in dP_m/dt_max the following must be taken into consideration: (a) that the measurement occurs only at the initiation of the inspiration and the shape of the respiratory muscle pressure generation throughout inspiration may be influenced by respiratory system unloading (Gallagher et al. 1989); (b) any changes in end-expiratory lung volume would alter inspiratory muscle length and the magnitude of pressure generated for any given neural activation of the muscle; (c) if expiration becomes active, abdominal pressure shifts in the negative direction during early inspiration and dP_m/dt_max will not represent predominantly the activity of inspiratory muscles (Grassino et al. 1981). Our second index, EMG_di, more closely represents 'neural' respiratory motor output, but also has at least two major potential confounding factors: (a) with increases in V_T during ACMV, it is likely that the proximity of the diaphragm EMG electrode to the muscle changed; and (b) large changes in diaphragmatic length have been shown to disrupt the relationship between EMG_di and phrenic nerve activity (Gandevia & McKenzie, 1986), although no dissociation between phrenic ENG and EMG_di were shown with more modest changes in diaphragm length during submaximal phrenic nerve stimulation (Road & Leevers, 1988).

We believe our experimental conditions minimized these potential confounding factors but we cannot be sure because we do not know, for example, how much diaphragm length actually changed during ACMV in sleeping humans. The resting (precontraction) length of the diaphragm appeared to be constant, since end-expiratory lung volumes were the same during ACMV and spontaneous breathing. The reductions in EMG_di and dP_m/dt_max obtained consistently among subjects and upon repeat trials of normocapnic ACMV in the same subject was encouraging evidence that both mechanical and electrical measures from the inspiratory muscles were at least changing in the same direction and to a similar extent. Most importantly, we found that changes in EMG_di and dP_m/dt_max were in the same direction as changes in V_T and V_T/T_I during the first few spontaneous breaths following ACMV. We believe this similarity in the direction of change in all three measurements to be especially important because the observed reductions in ventilatory output immediately following ACMV represent straightforward, accurately measurable reflections of respiratory motor output. This interpretation of changes in V_T or V_T/T_I between eupnoea (control) and post-ACMV assumes that airway resistance was not increased following ACMV. The constancy of upper airway resistance in non-snoring healthy subjects under these conditions has been previously reported (Badr et al. 1994).

These indirect indices were not able to discriminate small variations in respiratory motor output. For example, changes in dP_m/dt_max during ACMV accounted for only 40 % of the variation among subjects in EMG_di. Furthermore, when respiratory motor output was very low (which occurred immediately following ACMV), EMG_di was also markedly reduced but uncorrelated with reductions in V_T or V_T/T_I. Accordingly, we believe the correspondence among measures is sufficient to support the claim that normocapnic ACMV at raised V_T caused a significant, progressive, time-dependent reduction in the amplitude of inspiratory motor output and that this reduction was maintained during the initial recovery from ACMV. On the other hand, we do not think these indices permit a more precise quantification of small variations in respiratory motor output during and following ACMV, especially when inspiratory motor output appears to be very low.

To determine whether an inspiratory effort occurred for the purpose of measuring breath timing, we looked for a negative deflection in the mask pressure tracing accompanied by evidence of EMG_di activity. Weak inspiratory efforts were occasionally difficult to detect because of noise in either the mask pressure or EMG_di signal. Therefore, we cannot be completely sure that we were able to detect all weak inspiratory efforts.

Given the high sensitivity of breath timing and amplitude to even very small reductions in P_a,CO2 during NREM sleep (Skatrud & Dempsey, 1983; Henke et al. 1988; Datta et al. 1991; Meza et al. 1998), it was important that the concentration of CO₂ in the environment of the chemoreceptors was maintained during the 'normocapnic' trials of increased V_T via ACMV. It is conceivable that an increased F_I,CO2 and/or positive pressure mechanical ventilation may have altered the end-tidal to arterial P_CO2 difference. Our serial sampling of arterial blood showed, however, that the end-tidal to arterial P_CO2 difference was not altered during ACMV and these data are consistent with those of Simon et al. (1993) obtained during prolonged controlled mechanical ventilation. This evidence, along with the fact that P_ET,CO2 was, on average, held significantly greater than eupnoeic control values (+0·6 to 1·1 mmHg P_ET,CO2) for most 'normocapnic' ACMV trials (see Fig. 2A and B), demonstrates that the reduction in respiratory motor output observed during ACMV at raised V_T was, indeed, independent of reduced chemoreceptor stimuli.

Other potential causes of reduced respiratory motor output during normocapnic ACMV would include increased baroreceptor input and/or increased cerebral blood flow, with the latter causing hypocapnia and alkalosis in cerebral fluids and, therefore, inhibition of medullary chemoreceptors in the face of isocapnic arterial P_CO2. We have no direct measurements of either of these parameters; however, we have observed that systemic arterial blood pressure remains unchanged from control conditions throughout positive pressure mechanical ventilation at high V_T (Seals et al. 1990).

Effects of ACMV at increased tidal volume on respiratory motor output

The major support for non-chemical inhibition of respiratory motor output found in our study was that EMG_di and dP_m/dt_max were reduced as V_T increased during ACMV and P_a,CO2 was held above normocapnic levels. This inhibitory effect was significant with as little as a 28 % increase in V_T and continued even after ACMV was terminated. Breathing frequency fell significantly during normocapnic trials only at the higher V_T, but T_tot was only 26 % longer than control T_tot. Thus, apnoea was not elicited by ACMV at high V_T alone.

Our study focused on the time course effects of an increased V_T during ACMV in sleep and so the experimental design differed from that of other reported studies which have also used mechanical ventilation. However, there are some reports which support our findings of a non-chemical effect on respiratory motor output in sleeping humans. Henke et al. (1988) and Manchanda et al. (1996) used several minutes of normocapnic mechanical ventilation with small increases in V_T (less than 20 % > eupnoeic value) and ventilator frequency set at eupnoeic levels and caused consistent reductions in EMG_di amplitude. Similarly, the inspiratory pressure support mode of mechanical ventilation also caused immediate reductions in EMG_di in sleeping humans (Morrell et al. 1993) (see below). Other studies of mechanical ventilation-induced increases in V_T during sleep with maintained normocapnia reported moderate but significant slowing of breathing frequency which was similar to our findings, but changes in the amplitude of respiratory motor output during controlled mechanical ventilation were not determined (Leevers et al. 1994; Tobert et al. 1997).

Analysis of time course effects, both during normocapnic ACMV and in recovery, showed a clear distinction between the effects of ACMV on amplitude versus timing of respiratory motor output. A time-dependent effect of normocapnic ACMV at high V_T on the amplitude of respiratory motor output was indicated by the progressive fall in EMG_di and dP_m/dt_max over the first several ventilator cycles. Furthermore, this inhibitory effect persisted following removal of the ACMV since all three indices of respiratory motor output (including spontaneous V_T and V_T/T_I) remained reduced, and gradually recovered over a few spontaneous breaths. Apparently, then, the feedback effects of mechanical ventilation are not only cumulative with repeated cycles of ACMV, but also show an 'inertial' or inhibitory 'memory' effect once the sensory mechanical feedback was removed. While these carryover effects have been shown following passive mechanical ventilation (i.e. a persistent apnoea occurs) (Leevers et al. 1994), this is the first demonstration we know of a persistent inhibitory effect on respiratory motor output following mechanical ventilation wherein each ventilator cycle was actively initiated by the subject and hypocapnia was not present.

In contrast, ACMV effects (even at high V_T) on T_E and T_tot were small and occurred on the first ACMV cycle with little change thereafter. Upon cessation of ACMV, T_tot returned to even less than control on the first recovery spontaneous breath. These effects of ACMV on the timing of respiratory motor output were, then, purely a function of the mechanical characteristics of the machine breath, with no evidence of either cumulative or carryover inhibitory effects (see also the next section).

The reductions in respiratory motor output during positive pressure ACMV at raised V_T may be due to one or more of several types of mechanical perturbation. In addition to the increased V_T, positive pressure mechanical ventilation also 'unloaded' inspiratory muscles. Pressure support per se has been shown to cause downregulation of respiratory motor output, although accompanying transient changes in P_ET,CO2 often complicate interpretation of these changes (Lofaso et al. 1992; Morrell et al. 1993; Meza et al. 1998). This unloading effect may even be enhanced in NREM sleep during which upper airway resistance is commonly increased. However, our subjects were all non-snorers with no evidence of inspiratory flow limitation during sleep; therefore, we would not expect substantial changes in airway resistance during mechanical ventilation (Henke et al. 1988; Badr et al. 1994; Morrell et al. 1995). Nevertheless, we were unable to distinguish between any purely pressure support effects of ACMV and the influence of an increasing V_T.

Effects of changing flow rate and T_I

Although our aim was to increase V_T, the use of ACMV for this purpose required that inspiratory flow rate and T_I were also manipulated; how one controls these latter two parameters will also affect timing. We increased peak inspiratory flow rate by 1·7 times eupnoea in order to achieve the highest V_T. Increases in inspiratory flow rate in this range have been shown to shorten T_tot less than 10 % during sleep, with no effect on the amplitude of respiratory motor output (i.e. dP_m/dt_max) (Georgopoulos et al. 1996). In order to minimize increases in flow rate to this range, we maintained a T_I which was about 30 % longer than eupnoeic control at the higher normocapnic V_T during ACMV (see Table 1). This means that lung inflation occurred during a portion of the normal expiratory phase of the breath which, if the volume is sufficiently high, will probably delay the onset of the next inspiratory effort (Younes & Polacheck, 1985). This small prolongation of T_E with increased T_I at high V_T may occur because of vagal feedback from lung stretch. This was shown by the classic Hering-Breuer inhibitory reflex obtained during held inflations of greater than 1 l in sleeping humans (Hamilton et al. 1988) which was not present in lung denervated transplant patients (Iber et al. 1995). On the other hand, Tobert et al. (1997) reported that lung transplant patients also showed a slowing of breathing frequency in response to prolonged ventilator T_I at raised V_T.

Effects of increased V_T versus hypocapnia

In order to cause apnoea with ACMV, the P_ET,CO2 had to be reduced to at least 3 mmHg less than the eupnoeic P_ET,CO2 in NREM sleep and this hypocapnia had to be sustained for an average of 25-30 s or six to seven ventilator cycles at increased V_T. The difference in findings between normocapnic and hypocapnic ACMV does not mean that hypocapnia has a primary effect on breath timing mechanisms whereas elevated V_T (or unloading) affects only the amplitude of respiratory motor output. Rather, it is more likely that both the mechanical and chemical effects are primarily on the amplitude of respiratory motor output and that hypocapnia simply caused more complete inhibition of amplitude, thereby delaying the onset of the next inspiratory effort and prolonging T_E.

This effect of hypocapnia on prolonging T_E during NREM sleep is consistent with the concept of a sensitive hypocapnia-induced apnoeic threshold during sleep (Fink et al. 1963; Skatrud & Dempsey, 1983; Henke et al. 1988; Datta et al. 1991; Meza et al. 1998). However, a puzzling finding was the marked variability in T_E prolongation obtained during ACMV trials at any given reduction in P_ET,CO2 and, especially, the relatively minor changes in breath timing which sometimes accompanied even substantial levels of hypocapnia (i.e. > -5 mmHg P_ET,CO2) (see Fig. 5). Although we were unable to explain this heterogeneity in findings (see Results), we do note that these findings with relatively brief periods of progressive hypocapnia during ACMV differed from the much more consistent intra- and intersubject effects on the apnoeic threshold obtained using several minutes of varying levels of hypocapnia achieved by controlled mechanical ventilation (Skatrud & Dempsey, 1983; Henke et al. 1988; Datta et al. 1991). We believe that these differences point to the need for systematic investigation of the hypocapnia-induced apnoeic threshold, including the effect of hypocapnia duration, the effect of background (baseline) P_ET,CO2 (Xie et al. 1994) and the relative contributions of central versus peripheral chemoreceptor inhibition (Smith et al. 1995). Experimental testing of these influences is also important to determine exactly what conditions must be present in order for naturally occurring hypocapnia - especially transient hypocapnia - to cause central apnoea in sleeping humans.

Our present findings do not mean that non-chemical influences cannot exert profound effects on breath timing and even cause apnoeas. Nor do these findings mean that increased V_T, per se, does not have a major influence on apnoea duration. On the contrary, during several minutes of normocapnic controlled mechanical ventilation in NREM sleep, small increases in ventilator frequency above eupnoeic breathing frequency were previously shown to cause mechanical ventilation to become passive, and significant apnoeas to ensue upon cessation of mechanical ventilation (Leevers et al. 1994). What is apparently required to elicit these timing changes is an increase in ventilator frequency above the intrinsic rate which causes a resetting of breathing pattern with each ventilator cycle (Leevers et al. 1994). This mechanism of 'supercedence of spontaneous rhythm' for achieving passive conditions during mechanical ventilation was recently suggested by Puddy et al. (1996). Furthermore, the findings of Knox (1973) in anaesthetized cats and our recent preliminary results in sleeping dogs (Satoh et al. 1998) illustrate how even low ventilator volumes delivered during early expiration will prolong T_E and, when repeated, will lead quickly to on-going passive mechanical ventilation while normocapnia is maintained. Given this background of rhythm resetting and passivity caused by increased frequency of mechanical ventilation, the magnitude of the imposed ventilator V_T (or positive pressure unloading effect) was shown to be an important determinant of the duration of the post ventilator apnoea in sleeping humans (Leevers et al. 1994).

Summary and physiological relevance

Our study used assist-control mechanical ventilation to show that sleeping humans experience a time-dependent inhibition of the amplitude of respiratory motor output, with little slowing of breathing frequency in response to positive pressure-induced increases in tidal volume without requiring an accompanying hypocapnia. Respiratory motor output amplitude also remained diminished immediately following cessation of normocapnic ACMV. These findings pertain to the regulation of breathing during mechanical ventilation. However, we were unable to distinguish between the effects of raised V_T versus muscle unloading, and it remains to be determined if a series of naturally augmented tidal volumes, per se, in sleeping humans will actually contribute to inhibition of respiratory motor output once the stimuli have been removed. Hypocapnia was required to cause substantial apnoea when V_T was raised using ACMV.

	REFERENCES

Top Abstract Introduction Methods Results Discussion References

Badr, M. S., Skatrud, J. B. & Dempsey, J. A. (1994). Effect of chemoreceptor stimulation and inhibition on total pulmonary resistance in humans during NREM sleep. Journal of Applied Physiology 76, 1682-1692	[Medline]
Corfield, D., Morrell, M. J. & Guz, A. (1995). The nature of breathing during hypocapnia in awake man. Respiration Physiology 101, 145-160	[Medline]
Datta, A. K., Shea, S. A., Horner, R. L. & Guz, A. (1991). The influence of induced hypocapnia and sleep on the endogenous respiratory rhythm in humans. The Journal of Physiology 440, 17-33	[Abstract]
Fink, B. R., Hanks, E. C., Ngal, E. H. & Pappen, E. M. (1963). Central regulation of respiration during anesthesia and wakefulness. Annals of the New York Academy of Sciences 109, 892-899.
Fowler, W. S. (1954). Breaking point of breath-holding. Journal of Applied Physiology 6, 539-545.
Gallagher, C. G., Sunii, R. & Younes, M. (1989). Response of normal subjects to inspiratory resistive unloading. Journal of Applied Physiology 66, 1113-1119	[Medline]
Gandevia, S. C. & Mckenzie, D. K. (1986). Human diaphragmatic EMG: changes with lung volume and posture during supramaximal phrenic stimulation. Journal of Applied Physiology 60, 1420-1428	[Medline]
Georgopoulos, D., Mitrouska, I., Bshouty, Z., Anthonissen, N. R. & Younes, M. (1996). Effects of non-REM sleep on the response of respiratory output to varying inspiratory flow. American Journal of Respiratory and Critical Care Medicine 53, 1624-1630.
Georgopoulos, D., Mitrouska, I., Webster, K., Bshouty, Z. & Younes, M. (1997). Effect of inspiratory muscle unloading on the response of respiratory motor output to CO2. American Journal of Respiratory and Critical Care Medicine 155, 2000-2009	[Abstract]
Grassino, A. E., Derenne, J.-P., Almirall, J., Milic-Emili, J. & Whitelaw, W. (1981). Configuration of the chest wall and occlusion pressures in awake humans. Journal of Applied Physiology 50, 134-142	[Medline]
Hamilton, R. D., Winning, A. J., Horner, R. L. & Guz, A. (1988). The effect of lung inflation on breathing in man during wakefulness and sleep. Respiration Physiology 73, 145-154	[Medline]
Henke, K. G., Arias, A., Skatrud, J. B. & Dempsey, J. A. (1988). Inhibition of inspiratory muscle activity during sleep: chemical and non-chemical influences. American Review of Respiratory Disease 138, 8-15	[Medline]
Iber, C., Simon, P. M., Skatrud, J. B., Mahowald, M. W. & Dempsey, J. A. (1995). The Breuer-Hering reflex in humans: effects of pulmonary denervation and hypocapnia. American Journal of Respiratory and Critical Care Medicine 152, 217-224	[Abstract]
Knox, C. K. (1973). Characteristics of inflation and deflation reflexes during expiration in the cat. Journal of Neurophysiology 36, 284-295	[Medline]
Leevers, A. M., Simon, P. M. & Dempsey, J. A. (1994). Apnea after normocapnic mechanical ventilation during NREM sleep. Journal of Applied Physiology 77, 2079-2085	[Medline]
Lofaso, F., Isabey, D., Harf, A. & Scheid, P. (1992). Respiratory response to positive and negative inspiratory pressure in humans. Respiration Physiology 89, 75-88	[Medline]
Manchanda, S., Leevers, A. M., Wilson, C. R., Simon, P. M., Skatrud, J. B. & Dempsey, J. A. (1996). Frequency and volume thresholds for inhibition of inspiratory motor output during mechanical ventilation. Respiration Physiology 105, 1-16	[Medline]
Matthews, A. W. & Howell, J. B. L. (1975). The rate of isometric inspiratory pressure development as a measure of responsiveness to CO2 in man. Clinical Sciences and Molecular Medicine 49, 57-68.
Meza, S., Mendez, M., Ostrowski, M. & Younes, M. (1998). Determinants of periodic breathing (PB) with pressure support ventilation (PSV) during sleep. American Journal of Respiratory and Critical Care Medicine 157, A779.
Morrell, M. J., Harty, H. R., Adams, L. & Guz, A. (1995). Changes in total pulmonary resistance and PCO2 between wakefulness and sleep in normal human subjects. Journal of Applied Physiology 78, 1339-1349	[Medline]
Morrell, M. J., Shea, S. A., Adams, L. & Guz, A. (1993). Effect of inspiratory pressure support upon breathing in humans during wakefulness and sleep. Respiration Physiology 93, 57-70	[Medline]
Puddy, A., Patrick, W., Webster, K. & Younes, M. (1996). Respiratory control during volume-cycled ventilation in normal humans. Journal of Applied Physiology 80, 1749-1758	[Medline]
Road, J. D. & Leevers, A. M. (1988). Effect of lung inflation on diaphragmatic shortening. Journal of Applied Physiology 65, 2383-2389	[Medline]
Satoh, M., Eastwood, P. R., Curran, A. K., Smith, C. A. & Dempsey, J. A. (1998). Effect of mechanical ventilation applied at end-expiration on respiratory muscle activity in sleeping and anesthetized dogs. American Journal of Respiratory and Critical Care Medicine 157, A341.
Scheid, P., Lofaso, F., Isabey, D. & Harf, A. (1994). Respiratory response to inhaled CO2 during positive inspiratory pressure in humans. Journal of Applied Physiology 77, 876-882	[Medline]
Seals, D. R., Suwarno, N. O. & Dempsey, J. A. (1990). Influence of lung volume on sympathetic nerve discharge in humans. Circulation Research 67, 130-141	[Abstract]
Simon, P. M., Dempsey, J. A., Landry, D. M. & Skatrud, J. B. (1993). Effect of sleep on respiratory muscle activity during mechanical ventilation. American Review of Respiratory Disease 147, 32-37	[Medline]
Simon, P. M., Skatrud, J. B., Badr, M. S., Griffin, D., Iber, C. & Dempsey, J. A. (1992). Role of airway mechanoreceptors in the inhibition of inspiration during mechanical ventilation in humans. American Review of Respiratory Disease 144, 1033-1041.
Skatrud, J. B. & Dempsey, J. A. (1983). Interaction of sleep state and chemical stimuli in sustaining rhythmic ventilation. Journal of Applied Physiology 55, 813-822	[Medline]
Smith, C. A., Saupe, K. W., Henderson, K. S. & Dempsey, J. A. (1995). Ventilatory effects of specific carotid body hypocapnia in dogs during wakefulness and sleep. Journal of Applied Physiology 79, 689-699	[Medline]
Tobert, D. G., Simon, P. M., Stroetz, R. W. & Hubmayr, R. (1997). The determinants of respiratory rate during mechanical ventilation. American Journal of Respiratory and Critical Care Medicine 155, 485-492	[Abstract]
Xie, A., Wong, B., Phillipson, E. A., Slutsky, A. S. & Bradley, T. D. (1994). Interaction of hyperventilation and arousal in the pathogenesis of idiopathic central sleep apnea. American Journal of Respiratory and Critical Care Medicine 150, 489-495	[Abstract]
Younes, M. & Polacheck, J. (1985). Central adaptation to inspiratory inhibitory, expiratory prolonging vagal input. Journal of Applied Physiology 59, 1072-1084	[Medline]

Acknowledgements

The authors thank D. Puleo for technical assistance. This research was funded by a National Heart, Lung, and Blood Institute Specialized Centre of Research grant HL-42242 and by the Veterans Affairs Medical Research Service.

Corresponding author

C. R. Wilson: Department of Physical Therapy, 6RB, Northeastern University, 360 Huntington Avenue, Boston, MA 02115, USA.

Email: crwilson{at}lynx.neu.edu

This article has been cited by other articles:

				A. Xie, J. B. Skatrud, D. S. Puleo, and J. A. Dempsey Influence of arterial O2 on the susceptibility to posthyperventilation apnea during sleep J Appl Physiol, January 1, 2006; 100(1): 171 - 177. [Abstract] [Full Text] [PDF]

				S. Thomson, M. J. Morrell, J. J. Cordingley, and S. J. Semple Ventilation is unstable during drowsiness before sleep onset J Appl Physiol, November 1, 2005; 99(5): 2036 - 2044. [Abstract] [Full Text] [PDF]

				J. A. Dempsey Crossing the apnoeic threshold: causes and consequences Exp Physiol, January 1, 2005; 90(1): 13 - 24. [Abstract] [Full Text] [PDF]

				H. E. Cooper, T. H. Clutton-Brock, and M. J. Parkes Contribution of the respiratory rhythm to sinus arrhythmia in normal unanesthetized subjects during positive-pressure mechanical hyperventilation Am J Physiol Heart Circ Physiol, January 1, 2004; 286(1): H402 - H411. [Abstract] [Full Text] [PDF]

				T. Sharshar, G. Desmarais, B. Louis, G. Macadou, R. Porcher, A. Harf, J.-C. Raphael, D. Isabey, and F. Lofaso Transdiaphragmatic Pressure Control of Airway Pressure Support in Healthy Subjects Am. J. Respir. Crit. Care Med., October 1, 2003; 168(7): 760 - 769. [Abstract] [Full Text] [PDF]

				A. J. Rice, H. C. Nakayama, H. C. Haverkamp, D. F. Pegelow, J. B. Skatrud, and J. A. Dempsey Controlled versus Assisted Mechanical Ventilation Effects on Respiratory Motor Output in Sleeping Humans Am. J. Respir. Crit. Care Med., July 1, 2003; 168(1): 92 - 101. [Abstract] [Full Text] [PDF]

				S. Corne, K. Webster, and M. Younes Hypoxic Respiratory Response during Acute Stable Hypocapnia Am. J. Respir. Crit. Care Med., May 1, 2003; 167(9): 1193 - 1199. [Abstract] [Full Text] [PDF]

				M. Younes, A. Xie, J. B. Skatrud, C. A. Smith, and J. A. Dempsey Measurement of the co2 apneic threshold Am. J. Respir. Crit. Care Med., February 1, 2003; 167(3): 472 - 473. [Full Text] [PDF]

				S. S Sprague and P. D Hopkins Use of Inspiratory Strength Training to Wean Six Patients Who Were Ventilator-Dependent Physical Therapy, February 1, 2003; 83(2): 171 - 181. [Abstract] [Full Text] [PDF]

				A. Xie, J. B. Skatrud, D. S. Puleo, P. S. Rahko, and J. A. Dempsey Apnea-Hypopnea Threshold for CO2 in Patients with Congestive Heart Failure Am. J. Respir. Crit. Care Med., May 1, 2002; 165(9): 1245 - 1250. [Abstract] [Full Text] [PDF]