Brain hypoxia and control of breathing: neuromechanical control

The effects of graded brain hypoxia on respiratory cycle timing, the lung inflation reflex, and respiratory compensation for an inspiratory flow-resistive load were studied in unanesthetized goats. Two models, inhalation and CO and acute reduction of brain blood flow (BBF) were used to produce comparable levels of brain hypoxia. The lung inflation reflex was assessed as the ratio of inspiratory time of an occluded breath to that of the preceding spontaneous breath (TIoccl/TIspont). Compensation for flow-resistive loading was assessed as the effect of the load upon the airway occlusion pressure response to rebreathing CO2 (delta P 0.1/delta PCO2). Major findings were 1) severe brain hypoxia (HbCO of 60% or BBF of 42%) caused tachypnea due to a 50% or more reduction of expiratory time but only a 20% or less reduction of inspiratory time; 2) moderate carboxyhemoglobinemia (HbCO of 25-30%) enhanced TIoccl/TIspont from 1.5 +/- 0.1 at control to 2.1 +/- 0.1, while severe brain hypoxia (HbCO of 60% and BBF of 42%) reduced the ratio to 1.0 +/- 0.2; and 3) compensation for a flow-resistive load, manifested by increases of delta P 0.1/delta PCO2 of 75-300% in the control state, was abolished at HbCO of 45-50% and BBF of 60%. The data suggest that in unanesthetized animals brain hypoxia elicits tachypnea largely by an effect on the expiratory phase of the bulbopontine timing mechanism. The observed enhancement of the lung inflation reflex and abolition of flow-resistive load compensation are best explained by hypoxic depression of higher than brain stem neural function.     

Augmentation of CO2 drives by chlormadinone acetate, a synthetic progesterone

We studied 10 male subjects who were administered chlormadinone acetate (CMA), a potent synthetic progesterone, to clarify the physiological basis of its respiratory effects. Arterial blood gas tension, resting ventilation, and respiratory drive assessed by ventilatory and occlusion pressure response to CO2 with and without inspiratory flow-resistive loading were measured before and 4 wk after CMA administration. In all subjects, arterial PCO2 decreased significantly by 5.7 +/- 0.6 (SE) Torr with an increase in minute ventilation by 1.8 +/- 0.6 l X min-1, whereas no significant changes were seen in O2 uptake. During unloaded conditions, both slopes of occlusion pressure and ventilatory response to CO2 increased, being statistically significant in the former but showing nonsignificant trends in the latter. Furthermore, inspiratory flow-resistive loading (16 cmH2O X l(-1) X s) increased both slopes more markedly after CMA. The magnitudes of load compensation, assessed by the ratio of loaded to unloaded slope of the occlusion pressure response curve, were increased significantly. We concluded CMA is a potent respiratory stimulant that increases the CO2 chemosensitivity and neuromechanical drives in the load-compensation mechanism.     

Preterm infants: ventilation and P100 changes with CO2 and inspiratory resistive loading

The ventilatory effects of inspiratory flow-resistive loading and increased chemical drive were measured in ten neonates during progressive hypercapnia in control and loaded states. Hypercapnia (mean increase PCO2 = 15-20) resulted from inspiring 8% CO2 in room air and inspiratory loading by a flow-resistive load = 100 cmH2O X l-1) X s. Hypercapnia produced an increase in group minute ventilation secondary to increasing tidal volumes and breathing frequencies. Loading shifted the minute ventilation-CO2 response to the right, and slopes decreased significantly (P less than 0.05) consequent to a significant decrease in the frequency-CO2 slopes (P less than 0.05), which became negative in four of the ten subjects. Mouth pressure measured at 100 ms after onset of inspiratory effort (P100) occlusion pressure-CO2 slopes measured in five subjects showed no significant increase with load application. Resistive loading produced significant increases in inspiratory time (P less than 0.02) and the inspiratory time/total breath time ratio (P less than 0.01). Airway occlusion elicited the Hering-Breuer reflex, with a significant increase in inspiratory time-to-total breath time ratio (P less than 0.01). The results show that the inspiratory resistive load produced ventilatory compromise in newborns and insufficient compensatory augmentation of central drive.     

Background ventilatory stimulus as a determinant of load compensation

Compensation for inspiratory flow-resistive loading was compared during progressive hypercapnia and incremental exercise to determine the effect of changing the background ventilatory stimulus and to assess the influence of the interindividual variability of the unloaded CO2 response on evaluation of load compensation in normal subjects. During progressive hypercapnia, ventilatory response was incompletely defended with loading (mean unloaded delta VE/delta PCO2 = 3.02 +/- 2.29, loaded = 1.60 +/- 0.67 1.min-1.Torr-1 CO2, where VE is minute ventilation and PCO2 is CO2 partial pressure; P less than 0.01). Furthermore the degree of defense of ventilation with loading was inversely correlated with the magnitude of the unloaded CO2 response. During exercise, loading produced no depression in ventilatory response (mean delta VE/delta VCO2 unloaded = 20.5 +/- 1.9, loaded = 19.2 +/- 2.5 l.min-1.l-1.min-1 CO2 where VCO is CO2 production; P = NS), and no relationship was demonstrated between degree of defense of the exercise ventilatory response and the unloaded CO2 response. Differences in load compensation during CO2 rebreathing and exercise suggest the presence of independent ventilatory control mechanisms in these states. The type of background ventilatory stimulus should therefore be considered in load compensation assessment.     

Inspiratory resistive load detection in conscious dogs

The physiological mechanisms mediating the detection of mechanical loads are unknown. This is, in part, due to the lack of an animal model of load detection that could be used to investigate specific sensory systems. We used American Foxhounds with tracheal stomata to behaviorally condition the detection of inspiratory occlusion and graded resistive loads. The resistive loads were presented with a loading manifold connected to the inspiratory port of a non-rebreathing valve. The dogs signaled detection of the load by lifting their front paw off a lever. Inspiratory occlusion was used as the initial training stimulus, and the dogs could reliably respond within the first or second inspiratory effort to 100% of the occlusion presentations after 13 trials. Graded resistances that spanned the 50% detection threshold were then presented. The detection threshold resistances (delta R50) were 0.96 and 1.70 cmH2O.l-1.s. Ratios of delta R50 to background resistance were 0.15 and 0.30. The near-threshold resistive loads did not significantly change expired PCO2 or breathing patterns. These results demonstrate that dogs can be conditioned to reliably and specifically signal the detection of graded inspiratory mechanical loads. Inspiration through the tracheal stoma excludes afferents in the upper extrathoracic trachea, larynx, pharynx, nasal passages, and mouth from mediating load detection in these dogs. It is unknown which remaining afferents (vagal or respiratory muscle) are responsible for load detection.     

Diaphragm fatigue induced by inspiratory resistive loading in spontaneously breathing rabbits

Diaphragmatic contractility was assessed in spontaneously breathing ketamine-anesthetized rabbits by measuring the strength of diaphragmatic contraction in response to bilateral supramaximal phrenic nerve stimulation at frequencies between 10 and 100 Hz. During 10-180 min of inspiratory resistive loading, contractility decreased by approximately 40%, and hypoxemia and both respiratory and lactic acidosis developed. After 10 min of recovery, both the response to high-frequency stimulation (100 Hz) and the arterial PO2 and PCO2 returned to base-line levels, whereas metabolic acidosis and reduced response to low-frequency stimulation (10-20 Hz) persisted. Similar levels of hypoxemia and respiratory acidosis in the absence of inspiratory resistive loading did not alter diaphragmatic contractility. We conclude that in anesthetized rabbits excessive inspiratory resistive loading results in partially reversible diaphragm fatigue of the high- and low-frequency types, accompanied by hypoventilation and lactic acidosis.     

Immediate response to resistive loading in anesthetized humans

In eight spontaneously breathing anesthetized subjects (halothane: approximately 1 minimal alveolar concn; 70% N2O-30% O2), we determined 1) the inspiratory driving pressure by analysis of the pressure developed at the airway opening (Poao) during inspiratory efforts against airways occluded at end expiration; 2) the active inspiratory impedance; and 3) the immediate (first loaded breath) response to added inspiratory resistive loads (delta R). Based on these data we made model predictions of the immediate tidal volume response to delta R. Such predictions closely fitted the experimental results. The present investigation indicates that 1) in halothane-anesthetized humans the shape of the Poao wave differs from that in anesthetized animals, 2) the immediate response to delta R is not associated with appreciable changes in intensity, shape, and timing of inspiratory neural drive but depends mainly on intrinsic (nonneural) mechanisms; 3) the flow-dependent resistance of endotracheal tubes must be taken into account in studies dealing with increased neuromuscular drive in intubated subjects; and 4) in anesthetized humans Poao reflects the driving pressure available to produce the breathing movements.     

Naloxone alters the early response to an inspiratory flow-resistive load

In a previous study in unanesthetized goats, we demonstrated that cerebrospinal fluid levels of beta-endorphin were significantly elevated after 2.5 h of inspiratory flow-resistive loading. Naloxone (NLX) (0.1 mg/kg) administration partially and transiently reversed the tidal volume depression seen during loading. In the current study, we tested the hypothesis that endogenous opioid elaboration results in depression of respiratory output to the diaphragm. In six studies of five unanesthetized goats, tidal volume (VT), transdiaphragmatic pressure (Pdi), diaphragmatic electromyogram (EMGdi), and arterial blood gases were monitored. A continuous NLX (0.1 mg/kg) or saline (SAL) infusion was begun 5 min before an inspiratory flow-resistive load of 120 cmH2O.l-1.s was imposed. Our data show that the depression of VT induced by the load was prevented by NLX as early as 15 min and persisted for 2 h. At 2 h, Pdi was still 294 +/- 45% of the base-line value compared with 217 +/- 35% during SAL. There was no difference in EMGdi between the groups at any time. However, the augmentation of Pdi was associated with a greater increase in end-expiratory gastric pressure in the NLX group. We conclude that the reduction in VT and Pdi associated with endogenous opioid elaboration is not mediated by a decrease in neural output to the diaphragm, but it appears to be the result of a decrease in respiratory output to the abdominal muscles.     

Brain hypoxia preferentially stimulates genioglossal EMG responses to CO2

Although the dominant respiratory response to hypoxia is stimulation of breathing via the peripheral chemoreflex, brain hypoxia may inhibit respiration. We studied the effects of two levels of brain hypoxia without carotid body stimulation, produced by inhalation of CO, on ventilatory (VI) and genioglossal (EMGgg) and diaphragmatic (EMGdi) responses to CO2 rebreathing in awake, unanesthetized goats. Neither delta VI/delta PCO2 nor VI at a PCO2 of 60 Torr was significantly different between the three conditions studied (0%, 25%, and 50% carboxyhemoglobin, HbCO). There were also no significant changes in delta EMGdi/delta PCO2 or EMGdi at a PCO2 of 60 Torr during progressive brain hypoxia. In contrast, delta EMGgg/delta PCO2 and EMGgg at a PCO2 of 60 Torr were significantly increased at 50% HbCO compared with either normoxia or 25% HbCO (P less than 0.05). The PCO2 threshold at which inspiratory EMGgg appeared was also decreased at 50% HbCO (45.6 +/- 2.6 Torr) compared with normoxia (55.0 +/- 1.4 Torr, P less than 0.02) or 25% HbCO (53.4 +/- 1.6 Torr, P less than 0.02). We conclude that moderate brain hypoxia (50% HbCO) in awake, unanesthetized animals results in disproportionate augmentation of EMGgg relative to EMGdi during CO2 rebreathing. This finding is most likely due to hypoxic cortical depression with consequent withdrawal of tonic inhibition of hypoglossal inspiratory activity.     

Effects of taurine and a taurine antagonist on some respiratory and cardiovascuar parameters

Respiratory performance, heart rate and blood pressure were studied in halothane anesthetized rats after administration of taurine and the putative taurine antagonist 6-aminomethyl-3-methyl-4H-1,2,4-benzothiadiazine-1, 1-dioxide hydrochloride (TAG). Intracerebroventricular (i.c.v.) taurine depressed ventilation due to decreased inspiratory neural drive and depression of respiratory timing mechanisms. I.c.v. administration of 1–100 μg TAG caused no changes in the respiratory adn circulatory parameters studied except at the highest dose interval where respiratory frequency and minute ventilation were depressed. The respiratory depression induced by taurine (0.2 mg) or β-alanine (1 mg) was antagonized by administration of TAG (100 μg). However, TAG did not antagonize the respiratory effects induced by i.c.v. glycine or γ-aminobutyric acid (GABA) in equipotent respiratory depressant doses. The decline in inspiratory neural drive as well as in “respiratory timing” after i.c.v. taurine was restituted toward control values by TAG. The hypotension and bradycardia induced by taurine were also antagonized by TAG. It is concluded that TAG seems to antagonize the depressant action of taurine and β-alanine but not of GABA and glycine on respiratory performance. TAG might also possess some partial agonist activity in higher doses.     

Respiratory elastic load compensation in anesthetized patients with kyphoscoliosis

To evaluate the effects of abnormal respiratory mechanics and neuromuscular drive on the various components of elastic load compensation, we studied 16 anesthetized patients with kyphoscoliosis whose mean passive and active respiratory elastance (Ers and E'rs, respectively), active respiratory resistance, and peak inspiratory occlusion pressure were, respectively, 89, 84, 100, and 37% greater and inspiratory duration (TI) 13% less than corresponding values in 13 anesthetized controls. Ers comprised approximately 66% of effective elastance (E*rs) in both groups. E'rs, reflecting the role of the force-length properties of the active inspiratory muscles in increasing the internal impedance, comprised 83.8 and 86.1% of E*rs in the kyphoscoliosis patients and controls, respectively (P less than 0.001). This demonstrates the influence of increased intrinsic elastance and resistance and decreased TI on tidal volume defense in kyphoscoliosis patients in the absence of vagal modulation. In some patients the difference between Ers and E*rs was substantial, despite an unchanged or even shortened TI, suggesting that the Hering-Breuer reflex may affect stability through ways other than altering TI (e.g., via graded volume-dependent "terminal inhibition"). Characteristics of elastic load compensation in anesthetized kyphoscoliosis patients are similar to those of anesthetized normal subjects.     

Effort sensation, chemoresponsiveness, and breathing pattern during inspiratory resistive loading.

Although inspiratory resistive loading (IRL) reduces the ventilatory response to CO2 (VE/PCO2) and increases the sensation of inspiratory effort (IES), there are few data about the converse situation: whether CO2 responsiveness influences sustained load compensation and whether awareness of respiratory effort modifies this behavior. We studied 12 normal men during CO2 rebreathing while free breathing and with a 10-cmH2O.l-1.s IRL and compared these data with 5 min of resting breathing with and without the IRL. Breathing pattern, end-tidal PCO2, IES, and mouth occlusion pressure (P0.1) were recorded. Free-breathing VE/PCO2 was inversely related to an index of effort perception (IES/VE; r = -0.63, P less than 0.05), and the reduction in VE/PCO2 produced by IRL was related to the initial free-breathing VE/PCO2 (r = 0.87, P less than 0.01). IRL produced variable increases in inspiratory duration (TI), IES, and P0.1 at rest, and the change in tidal volume correlated with both VE/PCO2 (r = 0.63, P less than 0.05) and IES/VE (r = -0.69, P less than 0.05), this latter index also predicting the changes in TI with loading (r = -0.83, P less than 0.01). These data suggest that in normal subjects perception of inspiratory effort can modify free-breathing CO2 responsiveness and is as important as CO2 sensitivity in determining the response to short-term resistive loading. Individuals with good perception choose a small-tidal volume and short-TI breathing pattern during loading, possibly to minimize the discomfort of breathing.     

Relationship between respiratory nerve and muscle activity and muscle force output.

To demonstrate the most satisfactory way of using electrical activities of respiratory nerves and muscles, activities of phrenic nerve and external intercostal muscle (ICM) and the airway pressure changes generated by respiratory muscle contraction were recorded in anesthetized cats during complete airway occlusion. Electrical activities were rectified, integrated and processed in terms of peak and average inspiratory rates per 0.1 s and of total activity per breath. Peak rate of phrenic nerve activity exhibited a high linear correlation (r = 0.974) with peak inspiratory pressure. Average phrenic rate showed a similar high correlation (r = 0.973). Peak rate of external ICM was linearly related to peak pressure but the correlation was less good (r = 0.915). Total phrenic activity per breath was too dependent upon inspiratory duration to be a satisfactory correlate (r = 0.674). In this experiment occlusion pressure was an index of muscle force generation and respiratory control system output. It is concluded that peak or average rates of phrenic activity provide an electrical index of output changes. On theoretical grounds, peak rate is probably better.     

Effects of mechanical loading and hypercapnia on inspiratory muscle EMG.

The electromyograms of the diaphragm and an external intercostal muscle were analyzed to see if the effects of hypercapnia on inspiratory muscle electrical activity could be distinguished from those of mechanical loading and to determine whether changes in inspiratory muscle electrical activity were a sueful measure of CO2 response during mechanical loading. Anesthetized dogs were studied: 1) during progressive hypercapnia without mechanical loading, 2) during flow-resistive and elastic loading at constant PCO2, and 3) during progressive hypercapnia and mechanical loading. Both mechanical loading and hypercapnia increased total inspiratory diaphragmatic and intercostal muscle electrical activity. However, inspiratory duration was increased by mechanical loads but reduced by hypercapnia. Because of these changes in inspiratory duration, the average rate of diaphragmatic electrical activity remained unaffected by mechanical loading before and after vagotomy but was increased by hypercapnia. In contrast, both hypercapnia and mechanical loading increased the average rate of intercostal muscle electrical activity. There was a greater increase in both total and average rate of intercostal muscle electrical activity during hypercapnia in the presence of mechanical loading than during unloaded breathing. However, the change in total and average rate of diaphragmatic electrical activity with PCO2 was unaffected by added mechanical loads. These results suggest that diaphragmatic but not intercostal muscle electrical activity can be used as an index of CO2 response even during mechanical loading.     

Blood flow in respiratory muscles during maximal exertion in ponies with laryngeal hemiplegia

The interactive effects of upper airway negative pressure and hypercapnia on the pattern of breathing were assessed in pentobarbital-anesthetized cats. At any given level of pressure in the upper airway, hypercapnia increased respiratory rate, reduced inspiratory time, and augmented tidal volume, inspiratory airflow, and the peak and rate of rise of diaphragm electrical activity. Conversely, at any given level of CO2, upper airway negative pressure decreased respiratory rate, prolonged inspiratory time, and depressed inspiratory airflow and diaphragm electromyogram (EMG) rate of rise. Application of negative pressure to the upper airway shifted the relationship between tidal volume and inspiratory time upward and rightward. The relationship between inspiratory and expiratory times, however, was linearly correlated over a wide range of chemical drives and levels of upper airway pressure. These results suggest that in the anesthetized cat upper airway negative pressure afferent inputs 1) interact in an additive fashion with hypercapnia to alter the pattern of breathing, 2) interact multiplicatively with CO2 to influence mean inspiratory airflow and diaphragm EMG rate of rise, 3) depress the generation of central inspiratory activity, 4) increase the time-dependent volume threshold for inspiratory termination, and 5) affect the ratio between inspiratory and expiratory times in a similar manner as alterations in PCO2.     

Effect of resistive loading on inspiratory work output in anesthetized cats

In five spontaneously breathing anesthetized cats, we determined the inspiratory elastic (Wel), resistive (Wres), and total (WI) mechanical work rates (power) during control and first loaded inspirations through graded linear resistances (delta R) by "Campbell diagrams" based on measurement of esophageal pressure. WI did not change with delta R's up to 0.31 cmH2O X ml-1 X s, the concomitant decrease in Wel being balanced by an increase in Wres. The stability of WI in the face of delta R's was due to the vagally mediated prolongation of inspiration and the intrinsic properties of the respiratory system and of the contracting inspiratory muscles. To assess the separate contributions of volume-related and flow-related intrinsic mechanisms to the stability of WI, we made model predictions of the immediate effects of delta R's on inspiratory mechanical work output based on measurements of inspiratory driving pressure waves and passive and active respiratory resistance and elastance on the same five cats. The results suggest that the intrinsic stability of WI in the face of delta R's is provided primarily by the active elastance.     

Respiratory changes in parasternal intercostal intramuscular pressure

In an attempt to obtain insight in the forces developed by the parasternal intercostal muscles during breathing, changes in parasternal intramuscular pressure (PIP) were measured in 14 supine anesthetized dogs using a microtransducer method. In six animals, during bilateral parasternal stimulation a linear relationship between contractile force exerted on the rib and PIP was demonstrated (r greater than 0.95). In eight animals, during quiet active inspiration, substantial (55 +/- 11.5 cmH2O) PIP was developed. During inspiratory resistive loading and airway occlusion the inspiratory rise in PIP increased in proportion to the inspiratory fall in pleural pressure (r = 0.82). Phrenicotomy and vagotomy resulted in an increase in the inspiratory rise in PIP of 21% and 99%, respectively. During passive deflation, when the parasternal intercostals were passively lengthened, large rises (320 +/- 221 cmH2O) in intramuscular pressure were observed. During passive inflation intramuscular pressure remained constant or even decreased slightly (-8 +/- 25 cmH2O) as expected on the basis of the passive shortening of the muscles. PIP thus invariably increased when tension increased either actively or passively. From PIP it is clear that the parasternals exert significant forces on the ribs during respiratory maneuvers.     

Respiratory load compensation in chronic airway obstruction

To assess respiratory neuromuscular function and load compensating ability in patients with chronic airway obstruction (CAO), we studied eight stable patients with irreversible airway obstruction during hyperoxic CO2 rebreathing with and without a 17 cmH2O X l-1 X s flow-resistive inspiratory load (IRL). Minute ventilation (VE), transdiaphragmatic pressure (Pdi), and diaphragmatic electromyogram (EMGdi) were monitored. Pdi and EMGdi were obtained via a single gastroesophageal catheter with EMGdi being quantitated as the average rate of rise of inspiratory (moving average) activity. Based on the effects of IRL on the Pdi response to CO2 [delta Pdi/delta arterial CO2 tension (PaCO2)] and the change in Pdi for a given change in EMGdi (delta Pdi/delta EMGdi) during rebreathing, two groups could be clearly identified. Four patients (group A) were able to increase delta Pdi/delta PaCO2 and delta Pdi/delta EMGdi, whereas in the other four (group B) the IRL responses decreased. All group B patients were hyperinflated having significantly greater functional residual capacity (FRC) and residual volume than group A. In addition the IRL induced percent change in delta Pdi/delta PaCO2, and delta VE/delta PaCO2 was negatively correlated with lung volume so that in the hyperinflated group B the higher the FRC the greater was the decrease in Pdi response due to IRL. In both groups the greater the FRC the greater was the decrease in the ventilatory response to loading. Patients with CAO, even with severe airways obstruction, can effect load compensation by increasing diaphragmatic force output, but the presence of increased lung volume with the associated shortened diaphragm prevents such load compensation.     

Effects of changes in CO2 partial pressure on the sensation of respiratory drive

The purpose of this study was to determine whether a change in respiratory sensation accompanies an increase in CO2 partial pressure (PCO2) in the absence of any changes in the level and pattern of thoracic displacement and respiratory muscle force. Eleven normal subjects were artificially hyperventilated with a positive-pressure mechanical respirator. In separate trials the tidal volume (VT) was set at 10 and 18 ml/kg and the frequency of ventilation (f) was adjusted to maintain the base-line end-tidal PCO2 at approximately 30 Torr. Thereafter, at a constant controlled VT and f, the PCO2 was progressively increased by raising the inspired CO2 concentration. There were no changes in respiratory motor activity as determined from the peak inspiratory airway pressure (Paw) until the PCO2 reached 40.8 +/- 1.0 and 40.1 +/- 1.0 (SE) Torr in the large and small VT trials, respectively. Initially there was no conscious awareness of the change in respiratory activity. Subjects first signaled that ventilatory needs were not being satisfied only after a further increase in PCO2 to 44.7 +/- 1.3 and 42.3 +/- 1.0 (SE) Torr in the large and small VT trials and after the Paw had fallen to 55-60% of the base-line value. The results suggest that changes in respiratory sensation produced by increasing chemical drive are a consequence of increases in respiratory efferent activity, but a direct effect of changes in PCO2 on respiratory sensation cannot be excluded.     

Load compensation and respiratory muscle function during sleep.

The sleeping state places unique demands on the ventilatory control system. The sleep-induced increase in airway resistance, the loss of consciousness, and the need to maintain the sleeping state without frequent arousals require the presence of complex compensatory mechanisms. The increase in upper airway resistance during sleep represents the major effect of sleep on ventilatory control. This occurs because of a loss of muscle activity, which narrows the airway and also makes it more susceptible to collapse in response to the intraluminal pressure generated by other inspiratory muscles. The magnitude and timing of the drive to upper airway vs. other inspiratory pump muscles determine the level of resistance and can lead to inspiratory flow limitation and complete upper airway occlusion. The fall in ventilation with this mechanical load is not prevented, as it is in the awake state, because of the absence of immediate compensatory responses during sleep. However, during sleep, compensatory mechanisms are activated that tend to return ventilation toward control levels if the load is maintained. Upper airway protective reflexes, intrinsic properties of the chest wall, muscle length-compensating reflexes, and most importantly chemoresponsiveness of both upper airway and inspiratory pump muscles are all present during sleep to minimize the adverse effect of loading on ventilation. In non-rapid-eye-movement sleep, the high mechanical impedance combined with incomplete load compensation causes an increase in arterial PCO2 and augmented respiratory muscle activity. Phasic rapid-eye-movement sleep, however, interferes further with effective load compensation, primarily by its selective inhibitory effects on the phasic activation of postural muscles of the chest wall. The level and pattern of ventilation during sleep in health and disease states represent a compromise toward the ideal goal, which is to achieve maximum load compensation and meet the demand for chemical homeostasis while maintaining sleep state.