Control of ventilation

INTRODUCTION — The respiratory system is dependent upon adequate ventilation to supply oxygen, remove carbon dioxide, and help maintain acid-base homeostasis. Ventilation responds to changes in the arterial carbon dioxide tension (PaCO₂), arterial oxygen tension (PaO₂), and pH (figure 1), and may be modified in response to a number of mechanical and irritant stimuli arising from various structures within the thoracic cage, and probably from within muscles and joints during exercise.

Broadly viewed, the respiratory control mechanisms respond to input from neural and chemical receptors. Respiratory centers in the brain integrate these inputs and provide neuronal drive to the respiratory muscles, which maintain upper airway patency and drive the thoracic bellows to determine the level of ventilation [1,2].

The physiologic aspects of the control of ventilation and the evaluation of patients with disorders of ventilation will be reviewed here. The deleterious effects of various disease states on the control of ventilation are discussed separately. (See "Disorders of ventilatory control".)

RECEPTORS FOR VENTILATORY CONTROL

Thoracic neural receptors — Several different neural receptors are present in the upper airways, trachea, lungs, chest wall, and pulmonary vessels [3].

●Slowly adapting pulmonary stretch receptors and muscle spindles respond primarily to changes in lung volume.

●Rapidly adapting irritant receptors respond both to changes in lung volume and to the presence of chemicals such as histamine, prostaglandins, and exogenous noxious agents. The C-fiber (ie, small diameter, unmyelinated sensory afferent fibers) endings in the airways and lung respond primarily to their local chemical environment.

Activation of these receptors and fibers signals the respiratory centers via the vagus nerve and affects the breathing pattern by increasing the respiratory rate and/or stimulating cough, bronchoconstriction, and mucus production.

Input from these neural receptors likely accounts for the hyperventilation and hypocapnia that can occur in patients with pulmonary fibrosis even when hypoxemia is reversed by the administration of oxygen [4]. Hyperventilation may occur by this mechanism in patients with such problems as asthma, interstitial lung disease, pulmonary edema, pneumonia, and pulmonary embolism.

Peripheral chemoreceptors — The peripheral chemoreceptors, including the carotid and aortic bodies, are the primary sites for sensing the partial pressure of arterial oxygen (PaO₂), but they also increase their discharge in response to hypercapnia or acidosis. The carotid chemoreceptors are more important in adults; aortic chemoreceptors are active in infancy and childhood and then become relatively quiescent [5]. Neural outflow from the carotid bodies is estimated to account for up to 15 percent of resting ventilation, since peripheral chemodenervation leads to a rise in partial pressure of arterial carbon dioxide (PCO₂) of 5 to 10 mmHg [6,7].

The carotid bodies are located at the bifurcation of the common carotid arteries. Carotid body neuronal discharge increases as PaO₂ falls below 75 mmHg and becomes marked and progressive when PaO₂ is less than 50 to 55 mmHg [8]. The response of the carotid bodies to the combination of hypoxemia and hypercapnia is greater than the sum of the individual responses to each component. Carotid body resection leads to reduced hypoxic ventilatory response and to hypercapnia. (See "Disorders of ventilatory control", section on 'Cheyne-stokes respiration'.)

S-Nitrosothiols (SNOs) are thought to have a crucial role in signaling in the respiratory system including hypoxic ventilator drive [9]. Hypoxia leads to S-nitrosylation of Cystine thiols of certain proteins and peptides in red blood cells leading to S-nitrosothiols, which then activate carotid body cells. Impulses from the carotid bodies travel through the IX cranial nerve to the nucleus tractus solitarius (NTS), where excitatory neurotransmitters are released that result in increased ventilation [10]. Systemic N-acetylcysteine increases hypoxic ventilatory drive, likely acting via its effect on hemoglobin SNO [11].

There may be other, undefined peripheral receptors which help regulate ventilation during conditions of increased metabolic demand. With exercise, for example, there can be a 10-fold increase in the rate of metabolic carbon dioxide (CO₂) production, but arterial PCO₂ and pH are maintained at resting values until the anaerobic threshold is reached (see "Exercise physiology"). The increased ventilatory response to exercise occurs sooner than can be accounted for by changes in central PCO₂ [12]. The pH in the extracellular fluid of exercising muscle may be an important factor in increasing ventilation during aerobic exercise [13].

Further support for the existence of additional peripheral chemoreceptors comes from patients with congenital central hypoventilation (see "Disorders of ventilatory control", section on 'Congenital central hypoventilation syndrome' and "Congenital central hypoventilation syndrome and other causes of sleep-related hypoventilation in children", section on 'Congenital central hypoventilation syndrome'). These patients do not increase their ventilation when breathing a hypercapnic gas mixture; however, they increase their ventilation and maintain their pre-existing PCO₂ levels during exercise, and have a fall in PCO₂ during passive leg cycling [14,15].

Central chemoreceptors — Central nervous system (CNS) chemoreceptors have the major role in adjusting ventilation to maintain acid-base homeostasis. These receptors respond vigorously and almost instantaneously to changes in pH of the CNS environment. Several chemoreceptor sites have been identified in the medulla and mid-brain, the most significant ones being near the ventral surface of the medulla (VMS) and near the retrotrapezoid nucleus (RTN). RTN neurons receive input from carotid body chemoreceptors, and also express high levels of Phox2b, mutations of which are the cause of congenital central hypoventilation syndrome (CCHS) [16].

Acetylcholine and the parasympathetic nervous system play a major role determining the response to CO₂ stimulation, and probably contribute to normal central rhythm generation [10]. Congenital central hypoventilation syndrome has nearly absent ventilatory response to hypoxia and hypercapnia, is associated with Hirschsprung's disease, and usually requires ventilatory support. (See "Disorders of ventilatory control", section on 'Congenital central hypoventilation syndrome' and "Congenital central hypoventilation syndrome and other causes of sleep-related hypoventilation in children", section on 'Congenital central hypoventilation syndrome'.)

Because CO₂ is lipid-soluble, it crosses the blood-brain barrier rapidly; thus, changes in arterial carbon dioxide tension (PaCO₂) are sensed rapidly in the brain, producing changes in pH and ventilation. In comparison, the electrolyte composition in the CNS changes over a period of hours [17]. Thus, the effect on central chemoreceptors is less rapid with metabolic than with respiratory acid-base changes.

Ventral medullary pH and alpha-imidazole receptors — Chemoreceptors responsive to pH changes are scattered throughout the ventral surface of the medulla [18]. When these areas become more acidic, hyperventilation ensues and PaCO₂ is reduced. In contrast, when the pH rises at these chemoreceptor sites, hypoventilation occurs in order to raise partial pressure of arterial carbon dioxide (PCO₂) and normalize pH.

Ventilation is dependent upon brain interstitial fluid pH at constant body temperature [19]. However, ventilation does not track with pH if temperature varies; in this setting, ventilation correlates with the local levels of alpha-imidazole, the fractional dissociation of the imidazole moiety of histidine [20]. At a given temperature, small changes in brain interstitial fluid pH result in large changes in ventilation. For example, perfusion of the brain ventricular system of the turtle with mock cerebrospinal fluid (CSF) with pH 0.03 units below normal resulted in a four-fold increase in ventilation at 20ºC. A 2ºC rise in temperature also changes CSF pH by 0.03 units, but affects neither ventilation nor alpha-imidazole.

Regulation of respiration to keep a constant level of alpha-imidazole (the "alphastat") maintains relatively constant protein charge states and enzymatic functions despite changes in temperature and pH [21]. Alpha-imidazole regulation of ventilation is presumed to act at the ventral surface of the medulla [22,23].

Brain extracellular fluid composition — In addition to local PaCO₂, the pH of brain extracellular fluid (ECF) depends upon CSF electrolyte composition, which changes within several hours of arterial acid-base changes. Local regulation of brain ECF acid-base status is a complex process which involves a number of metabolically active cells, including glial cells, cells of the choroid plexus, and cells which maintain the brain blood-barrier [24].

Alterations in serum pH also eventually lead to changes in brain CSF composition, likely via effects on CSF formation by the choroid plexus. Chronic respiratory acidosis leads to a higher serum and CSF bicarbonate concentration, which ultimately lowers central ventilatory drive as the acidosis is ameliorated [25]. Similarly, CSF pH becomes alkaline due to chronic respiratory alkalosis, as occurs with exposure to high altitude; it then returns toward baseline over the next few days [26]. (See "High-altitude illness: Physiology, risk factors, and general prevention".)

NORMAL RESPONSES TO METABOLIC ACIDOSIS AND ALKALOSIS — There is a linear relationship between the plasma bicarbonate concentration and the change in arterial carbon dioxide tension (PaCO₂) in metabolic acidosis and alkalosis [27] (see "Simple and mixed acid-base disorders"). The expected compensations are somewhat different in these disorders:

●For every 1 mEq/L rise in plasma bicarbonate concentration in metabolic alkalosis, the PaCO₂ rises by 0.6 to 0.7 mmHg (figure 2) [27]. Thus, for a plasma bicarbonate concentration of 34 mEq/L (10 mEq/L above normal), the expected partial pressure of arterial carbon dioxide (PCO₂) is 46 to 47 mmHg.

●For every 1 mEq/L fall in plasma bicarbonate concentration in metabolic acidosis, the PaCO₂ falls by approximately 1.2 mmHg [28]. Thus, for a plasma bicarbonate concentration of 14 mEq/L (10 mEq/L below normal), the expected PCO₂ is 28 mmHg.

Values substantially different from expected reflect a superimposed acid-base disorder. (See "Simple and mixed acid-base disorders".)

INTEGRATION OF NEURAL AND CHEMORECEPTOR INPUT — Central respiratory centers receive stimulatory input from central respiratory pacer cells, central and peripheral chemoreceptors, upper airway receptors, other areas of the brain, and volitional pathways. Central respiratory centers integrate these signals into a combined output to the muscles of respiration. Much of the effect of peripheral chemoreceptors on ventilation appears to be mediated by its effect on central chemoreceptors. Carotid body peripheral chemoreceptor inputs (both O₂ and CO₂) have marked synergistic, hyperadditive influences on central chemoreceptor ventilatory sensitivity to central CO2 in studies in unanesthetized dogs [29,30]. During sleep, the respiratory drive is reduced, usually resulting in a rise in arterial carbon dioxide tension (PaCO₂) of 3 to 4 mmHg [31]. (See "Central sleep apnea: Pathogenesis".)

There normally is a mild inhibitory input from the cerebral cortex to the respiratory centers. Some stroke patients lose this inhibitory input leading to a lower baseline PaCO₂ and a tendency toward Cheyne-Stokes respiration [32]. There is also an increased tendency to develop obstructive sleep apnea related to decreased output to the upper airway muscles of respiration.

All of these stimulatory and inhibitory signals act to modify the characteristics of respiratory rhythm signals which emanate from central pattern generator cells in the medulla. From the medulla, neural outflow drives the breathing frequency, inspiratory time, and expiratory time.

Lesions of the brainstem may cause characteristic abnormalities in breathing pattern. Ataxic, or grossly irregular breathing can occur with medullary lesions, while apneustic breathing, characterized by sustained inspiration, can occur with pontine lesions.

EVALUATION OF RESPIRATORY DRIVE AND CONTROL — Assessment of a patient with hypercapnia or abnormal ventilation (eg, increased minute ventilation with a normal arterial carbon dioxide tension [PaCO₂]) should include clinical examination and pulmonary function testing to determine whether parenchymal, neuromuscular, or chest wall disease can explain the patient's abnormalities (table 1). (See "The evaluation, diagnosis, and treatment of the adult patient with acute hypercapnic respiratory failure".)

The use of pulmonary function testing to diagnose parenchymal lung disease is discussed separately. (See "Overview of pulmonary function testing in adults".)

The role of polysomnography in assessing changes in ventilatory control during sleep is also discussed separately. (See "Polysomnography in the evaluation of sleep-disordered breathing in adults" and "Disorders of ventilatory control" and "Central sleep apnea: Risk factors, clinical presentation, and diagnosis".)

Tests of respiratory control are primarily used for research purposes or under special circumstances when pulmonary function testing and respiratory muscle strength have failed to provide an explanation for abnormal levels of arterial oxygen tension (PaO₂) or PaCO₂. The various tests include measurement of hypoxic and hypercapnic ventilatory responses, mouth occlusion pressure, elastic and resistive load testing, and analysis of the patient's breathing pattern. Tests of ventilatory responses to hypoxia and hypercapnia are potentially hazardous, and there are significant variations among normals. As a result, the clinical condition of the patient and the indications for testing should be carefully considered. During testing, both pulse O₂ saturation and end-tidal CO₂ tension should be monitored.

Respiratory muscle forces — Respiratory muscle forces are easily measured with manometers to quantify airway pressure with the patient breathing against a closed mouthpiece. Maximum pressure generated at the beginning of a full inspiration from residual volume reflects the strength of the inspiratory muscles, while mouth pressure during a forceful expiration from total lung capacity reflects the strength of the expiratory muscles. Respiratory muscle force is dependent upon age, sex, and lung volume; all of these factors should be incorporated into the predicted values for each patient [33-35]. (See "Tests of respiratory muscle strength".)

Hypoxic challenge — Testing of the hypoxic ventilatory response is generally limited to research settings or rare circumstances when the etiology of hypoxemia is unclear. Testing is performed by delivering a hypoxic gas mixture (usually 10 to 12 percent O₂) to the patient via a rebreathing apparatus, resulting in progressive hypoxemia over four to six minutes. The partial pressures of alveolar and blood CO₂ are kept constant by removing CO₂ from the rebreathing bag with a soda lime apparatus [36]. Testing can also be performed by having the subject breathe a gas mixture with 8 percent O₂ for two minutes [37]. The hypoxic ventilatory response assesses the integrity of the peripheral chemoreceptors and the sensory pathways to the respiratory centers. As arterial oxygen tension falls, peripheral chemoreceptors are stimulated and respiratory drive increases.

The normal ventilatory response to oxygen desaturation is linear, with ventilation increasing approximately 1 L/min for each 1 percent fall in oxygen saturation (figure 3). The ventilatory response is curvilinear if plotted instead against PaO₂; as the PaO₂ falls below 50 to 55 mmHg, progressively larger increases in minute ventilation occur for equivalent reductions in oxygen tension. Most normal individuals increase their ventilation about three to six times resting values when the arterial PO₂ has reached a value of 40 mmHg, which generally corresponds to a hemoglobin saturation of 75 percent.

An abnormal hypoxic response generally indicates decreased output from the peripheral chemoreceptors, as occurs with various disease states or senescence, but is also observed in endurance athletes and in persons who reside at high altitudes [38-40]. In the latter two situations, modulation of the hypoxic ventilatory drive may be advantageous by limiting the sensation of dyspnea. (See "Exercise physiology" and "High-altitude illness: Physiology, risk factors, and general prevention".)

Hypercapnic challenge — Hypercapnic challenge testing is usually used in research settings. The ventilatory response to hypercapnia is usually determined by a rebreathing technique in which the subject rebreathes a hyperoxic gas mixture and gradually increases the PaCO₂ over four to six minutes [41,42]. A carbon dioxide analyzer in the mouthpiece measures the end-tidal CO₂ tension, which in normal lungs approximates the alveolar and arterial CO₂ tension. End-tidal partial pressure of arterial carbon dioxide (PCO₂) often poorly reflects PaCO₂ in patients with airflow obstruction and other lung abnormalities; direct measurement of arterial PaCO₂ is required in these patients.

The ventilatory response to CO₂ primarily reflects the activity of the central chemoreceptors. The rise in ventilation in response to increases in PaCO₂ is linear, generally in the range of 2.5 to 3 L/min for each mmHg increase in PaCO₂ [43].

Alterations in ventilatory response to CO2 and O2 — Approximately 15 percent of adults have a diminished response to CO₂, defined as an augmentation of ventilation of less than 1 L/min per mmHg rise in arterial PCO₂. These individuals are likely to develop CO₂ retention when additional respiratory problems arise, such as obesity, obstructive lung disease, or status asthmaticus [44-46] (see "Disorders of ventilatory control", section on 'Chronic obstructive pulmonary disease' and "Disorders of ventilatory control", section on 'Asthma'). As with reductions in hypoxic ventilatory drive, diminution of the response to hypercapnia is observed in elderly individuals, endurance athletes, and patients with hypothyroidism [38,47,48]. (See "Respiratory function in thyroid disease".)

Depressed ventilatory responses to CO₂ can also occur with elevation of serum and brain extracellular fluid (ECF) bicarbonate, such as occurs with metabolic alkalosis or chronic CO₂ retention [49]. With a higher brain ECF bicarbonate concentration, the pH will change less for a given change in PCO₂, producing less change in ventilation. The ventilatory response can be increased in this setting if the bicarbonate concentration is lowered. Thus, the slope of the ventilatory response to CO₂ needs to be evaluated in the context of the patient's acid-base status. (See "Pathogenesis of metabolic alkalosis".)

Patients with chronic CO₂ retention are often more dependent upon their hypoxic ventilatory drive because they are thought to have a depressed ventilatory response to CO₂ (due in large part to the compensatory elevation in the plasma bicarbonate concentration). In these individuals, administration of a high fraction of inspired O₂ can depress ventilation, particularly when the PaO₂ is raised above 60 mmHg. However, reduction in hypoxic vasoconstriction (with worsening of ventilation-perfusion mismatch) and a shift in the CO₂-hemoglobin dissociation curve (the Haldane effect) appear to be more important mechanisms for the increased CO₂ retention following O₂ administration [50]. Regardless of the mechanism, oxygen should not be withheld from hypoxic patients who have evidence of chronic CO₂ retention; in these circumstances oxygen is a life saving measure. (See "The evaluation, diagnosis, and treatment of the adult patient with acute hypercapnic respiratory failure".)

Native populations living at high or low altitude in different areas of the world have varied ventilatory responses. Tibetans have higher hypoxic ventilatory responses, higher resting ventilation, and higher resting ventilation variance compared with high altitude natives from the Andes and Rocky Mountains [51]. Distinct breathing patterns but similar acute hypoxic ventilator responses are found among genetically divergent low altitude populations [37].

Opioids are known to decrease the ventilatory response to hypercapnia. Whether other drugs commonly used in place of opioids can cause similar effects was studied in 25 healthy volunteers using carbon dioxide rebreathing technology [52]. Individuals were treated with oral quetiapine (twice daily; increasing daily doses from 100 to 400 mg) or paroxetine (40 mg daily) combined with oxycodone (10 mg daily) and compared with oxycodone alone. Paroxetine combined with oxycodone decreased the ventilatory response to hypercapnia, while quetiapine combined with oxycodone did not cause such an effect. Further study is needed to examine if this phenomenon is clinically relevant and persistent in the long term.

Elastic and resistive loading — Respiratory load compensation is another method of assessing ventilatory control mechanisms; this testing is typically performed in a research setting. Resistive load testing is done by having the subject breathe through progressively increasing resistance during inspiration or expiration. The increase in resistance is brought about by breathing through narrower and narrower tubes. The sensation of dyspnea perceived by the subject is scored by the Borg Scale from 0 to 10, with 0 being normal breathing and 10 being extreme dyspnea. Normal individuals have short expirations during elastic loading and long inspirations during inspiratory resistive loading.

Breathing pattern — The normal response to hypercapnic or hypoxic stimuli is to increase tidal volume with little change in respiratory frequency. Patients with a variety of neuromuscular diseases have a breathing pattern with a low tidal volume and a high frequency, and they maintain this pattern during hypoxic or hypercapnic stress [53,54]. Thus, a disproportionate rise in breathing frequency is suggestive of a disorder of respiratory muscle strength or a stimulus that arises from the thorax, rather than an abnormality of central control.

OTHER MEASUREMENTS OF RESPIRATORY EFFORT — The ventilatory drive of patients with severe pulmonary, chest wall, or muscle diseases may not be accurately reflected in measurements of minute ventilation. Respiratory muscle electromyograms or inspiratory pressure development during transient airway occlusion in response to the physiologic challenges listed above may be more useful parameters to follow in these patients.

Electromyography — Electromyography (EMG) permits a more direct measurement of neural output from the respiratory centers to the respiratory muscles than measurements of respiratory muscle strength, but is generally limited to the research setting. Such studies allow separate examination of intrinsic muscle strength and neural drive. As an example, surface electromyographic activity of diaphragmatic and intercostal muscles is increased in patients with myasthenia gravis compared with normal individuals, while respiratory muscle strength (eg, inspiratory and expiratory forces) is decreased [55]. Neural respiratory drive is also increased in patients with cystic fibrosis and chronic obstructive pulmonary disease (COPD) and is related to severity of disease [56,57].

The EMG response to CO₂ rebreathing does not increase as much as does the P0.1 response (see 'Mouth occlusion pressure' below). This suggests that there may be recruitment of additional respiratory muscles or enhanced diaphragm muscle efficiency during acute hypercapnia [58].

Mouth occlusion pressure — Measurement of the maximal pressure generated during the first 0.1 second of normal inspiratory effort when the mouth has been occluded (abbreviated P0.1) is helpful in evaluating central ventilatory drive. Effects of respiratory mechanics are minimized in the absence of airflow, and the occlusion is too short to be influenced by muscle weakness or by conscious alterations in respiration [59]. The lack of reduction in P0.1 with muscle weakness is illustrated by the fact that the measurement increases in normal volunteers given curare [60].

P0.1 can be used to assess the ventilatory response to hypoxia and hypercapnia and to assist with weaning patients from mechanical ventilation [61]. A high P0.1 indicates that the patient is exerting near maximal effort and has little reserve. (See "Weaning from mechanical ventilation: Readiness testing".)

P0.1 values decrease with age [62], and evidence suggests that individuals born with depressed P0.1 responses are predisposed to hypercapnia if they develop airways obstruction [44,45]. There is significant overlap in P0.1 between normal subjects and patients, but the ratio of minute ventilation to P0.1 can reliably differentiate between normal subjects and patients with lung disease. Patients with lung disease have a lower minute ventilation for a given P0.1 [63].

SUMMARY

●The respiratory system is dependent upon a variety of sensing (eg, peripheral stretch receptors and peripheral and central chemoreceptors) and control mechanisms to promote adequate ventilation to supply oxygen, remove carbon dioxide, and help maintain acid-base homeostasis. (See 'Introduction' above.)

●Respiratory centers in the medulla receive stimulatory input from central respiratory pacer cells, central and peripheral chemoreceptors, upper airway receptors, other areas of the brain, and volitional pathways and integrate these signals into a combined output to respiratory muscles to modulate breathing frequency, inspiratory time, and expiratory time. (See 'Integration of neural and chemoreceptor input' above.)

●The pH of brain extracellular fluid (ECF), as reflected in cerebrospinal fluid (CSF) electrolyte composition, changes within several hours of systemic acid-base changes. In metabolic alkalosis, for every 1 mEq/L rise in plasma bicarbonate concentration (HCO₃), the arterial carbon dioxide tension (PaCO₂) rises by 0.6 to 0.7 mmHg (figure 2). Thus, for a plasma bicarbonate concentration of 34 mEq/L (10 mEq/L above normal), the expected PaCO₂ is 46 to 47 mmHg. In metabolic acidosis, for every 1 mEq/L decrease in plasma bicarbonate concentration (HCO₃), the PaCO₂ falls by approximately 1.2 mmHg. Thus, for a plasma bicarbonate concentration of 14 mEq/L (10 mEq/L below normal), the expected PaCO₂ is 28 mmHg. (See 'Normal responses to metabolic acidosis and alkalosis' above and "Simple and mixed acid-base disorders".)

●The tests of ventilatory control are primarily used for research purposes or in rare situations when pulmonary function testing, respiratory muscle strength assessment, and polysomnography have failed to provide an explanation for abnormal levels of arterial oxygen tension (PaO₂) or PaCO₂. (See 'Evaluation of respiratory drive and control' above.)

●Assessment of ventilatory control may include measurement of hypoxic and hypercapnic ventilatory responses, elastic and resistive load testing, breathing pattern analysis, electromyography, and mouth occlusion pressure. (See 'Evaluation of respiratory drive and control' above.)

●Patients with congenital central hypoventilation syndrome and carotid body resection have abnormal ventilatory control. Some patients with chronic obstructive lung disease (COPD) and with asthma have abnormal ventilatory control. (See "Disorders of ventilatory control".)