Perspective on the human cough reflex

For my wife, it is reasonable to believe that instigation of a cough response is due to stimulation of a mechanically-sensitive 'true cough receptor' that is provoked by the sliver of ice. The premise for the presence of a 'true cough receptor' is explored by Canning et al using the anesthetized guinea pig model, which retains a blunted cough response following noxious stimuli [5, 6]. Mechanical stimulation, postnasal drip, and a water bolus placed into the pharynx will evoke coughing in both human subjects and in animals but not capsaicin [6, 7]. While C-fiber activation initiates coughing in conscious guinea pigs, it does not occur in the anesthetized animal. Anesthesia also attenuates or abolishes coughing in humans [8]. The anesthetized guinea pig model is important because it allows different investigative options, such as examining cough along with recordings from brainstem neurons or vagal afferent neurons, microinjection of drugs into specific brainstem and midbrain structures, selective stimulation of parts of the airways and not others, or extrinsic denervation of only parts of the airways [6]. My wife's coughing commenced with an inspiratory maneuver. In other cases, the mechanosensory-type reflex cough is associated with just a single and short expiratory cough referred to as an 'expiration reflex' [9‐11].

Purportedly, the vagal afferent neurons of the true cough receptor are unlike the rapidly adapting receptors (RARs), slowly adapting receptors (SARs) or C-fibers; also, this receptor does not express transient receptor potential vanilloid (TRPV1) and is not sensitive to capsaicin. In contrast to the anesthetized guinea pig model, mechanically-induced coughing in conscious guinea pigs generates impulses carried by low threshold, mechanically sensitive, rapidly adapting receptors (RARs) that travel through myelinated Aδ fibers at conduction velocities of between 4 and 18 m/s [12‐17]. These receptors also do not react directly to capsaicin and other chemical stimuli unless the stimulus leads to mechanical distortion of the nerve terminal [18]. RARs promote reflex bronchospasm and mucus secretion through parasympathetic pathways. How the conscious guinea pig cough mechanism applies to my wife's induced-coughing is not sufficiently understood.

Tracheal and laryngeal receptors come in to play as an important defensive role against acid aspiration [19]. The 'true cough receptor' is provoked by acid [6]. The descent of the larynx to a location more approximate to the opening of the esophagus places human ancestors at a greater risk for aspiration. Because of the greater risk for acid aspiration, possibly there is adaptation of a means to provoke coughing via a brainstem sensitizing mechanisms or by direct triggering of afferent esophageal nociceptors projecting from vagal pharyngeal and glossopharyngeal nerves [14, 20]. The ion channel receptor, TRPV1, respond to acid in a more sustained manner than the acid-sensing ion channel-type receptor (ASIC), which tends to produce brief and transient responses to a lowered pH [21].

Both the chemically- and mechanically-sensitive airway nerves take part in mediating the cough reflex and establishing synapses in the brainstem's caudal two-thirds of the nucleus tractus solitarius [22]. Because the threshold for mechanical activation is more conducive for RARs and SARs than for C-fibers type of nerves, C-fibers consequently respond less to mechanical stimuli than do RARs and SARs. In some animals (i.e., guinea pigs, rabbits, cats, dogs and pigs) glutamate may be the final central nervous system excitatory transmitter during coughing; neurokinins play a more modulatory role [23]. The chemosensory nociceptors reside as a fine plexus within the airway epithelium and walls and send nerve impulses through slowly conducting (< 2 m/s) vagal unmyelinated fibers [24‐28]. The sensation of an "urge to cough" is ostensibly associated with activation of bronchopulmonary C-fibers [29, 30]. C-fiber nerves become directly activated, 'sensitized' or 'hyper-activated' by capsaicin, bradykinin, adenosine, PGE₂, citric acid, hypertonic saline solution, SO₂ and lung inflammation (or inflammatory-type chemicals) [7, 18, 31‐33]. The role of lung neuropeptides and neuroinflammation in humans is poorly understood [34‐37].

C-fibers actions involve ion channels, ancient sensors of the environment. Hundreds of millions of years ago, ion channels first sense thermal and osmolality stimuli. Later, the ionic channel mechanisms are adapted for other 'environmental' sensations (e.g., hearing, vision and taste). TRPA₁, possibly an ''irritant receptor'', is expressed by a subpopulation of unmyelinated afferent C fiber nociceptors and may be linked to TRPV₁ to contribute to the transduction of the irritant stimuli [38‐43]. Mazzone questions the role of TRPV₁ and opines that it is improbable that any TRPV₁-expressing cells in other tissues or organs are involved in the cough reflex [18].

The motor pattern of the reflex cough is regulated differently than the motor pattern for tidal breathing [44]. Neuroplastic transformations allow respiratory behaviors having dissimilar spatial and temporal dimensions to instigate a mixture of actions employing comparable respiratory motoneurons [37, 45]. Animal investigations concerning the induction of Fos-like immunoreactive neurons in certain brain stem regions (i.e., the commissural subnucleus of the nucleus tractus solitarius, field of the reticular formation, and nucleus ambiguus) suggest that different reflex responses make use of overlapping neural circuitry [46]. The neural conduit for medullary control of laryngeal adductor muscles actions is contained within a broader neural pathways controlling cough and swallowing [46]. In animal models, there is activation of interneuron pathways located between the medullary nucleus tractus solitarius and the nucleus ambiguus during coughing [46, 47]. Stimulation of the superior laryngeal nerve can evoke different laryngeal adductor muscle responses including coughing, swallowing, gagging, laryngeal spasms, bronchoconstriction, apnea, and retching [46, 47]. The type of the evoked reflex response depends on the considerations of the stimulus used [46, 47].

While conceivably not evident to my wife, the motor side of her cough reflex is spaced by distinctive phases (Figure 1). During the inspiratory phase of cough, my wife's entire glottis is in abduction. The laryngeal motor neurons that begin in my wife's nucleus ambiguus follow vagal and superior laryngeal nerves to excite the motor neurons of her glottis, external intercostal muscles, diaphragm, and other major inspiratory and expiratory respiratory muscles [48]. Her upper airway motoneurons are located at the cranial level; phrenic motoneurons controlling the diaphragm are located in the cervical cord ventral horn; and, her rib cage and abdominal motoneurons are located in the ventral horn of the thoracolumbar segments [49, 50]. If measured, intracellular electrodes would record substantial depolarization of her laryngeal motor neurons during coughing. It seems that a central inspiratory command activates the respiratory motoneurons in an encoded sequential order; the upper airway motoneurons are recruited before those of the diaphragm and the rib cage; this allows the opening of the glottis before the fall in tracheal pressure due to diaphragmatic contraction [49]. The posterior cricoarytenoid muscle activates 40 to 100 milliseconds before the inspiratory activation of the diaphragm [51]. Actions taken by other inspiratory and the laryngeal abductor muscles (e.g., posterior cricoarytenoid muscle) further enlarge the opening of her upper airways [52‐60]. My wife's diaphragm and external intercostal muscles contract to expand her chest cavity and lower her intra-thoracic pressure. The crural and costal muscles of her diaphragm act in synchrony throughout inspiration [61]. The contraction of her diaphragm peaks within approximately 1.0 second. As her diaphragm descends, there is some widening of her glottal opening due to tautness placed on her larynx [62].

The compressive phase of her cough reflex begins almost immediately after inspiration. Her laryngeal motor neurons are transiently hyperpolarized during the transition between the inspiratory and expiratory compressive phases of her cough; laryngeal motor neurons are repolarized during the expiratory compressive phase [9]. The glottic closure is essential for the process of coughing since the maximal level of intrathoracic pressure attained, and the efficiency of the expiratory cough, depends in great part on the quality of glottic occlusion. The glottic closure reflex is elicited at birth, becomes more active during the first year of life, and gradually decreases in activity with further aging [51]. Very quickly and in the order of perhaps 200 milliseconds and a range between 42 and 1010 milliseconds, the glottic closure takes place as two small laryngeal abductors (posterior cricoarytenoid muscles) quickly relax and her laryngeal adductors (e.g., thyroarytenoid muscles) contract to produce significant narrowing or complete closure of the glottis [3, 63, 64]. The laryngeal sphincter muscles closure is so tight that it can sustain very high intratracheal pressures [65]. Intrapleural pressure may reach more than 100 cm H₂O [4, 56, 58, 65, 66]. There is coordinated activity between the posterior cricoarytenoid muscle and the accessory inspiratory muscles (i.e., infrahyoid and intercostal muscles) [51, 57, 67]. Synchronized actions are taken by her expiratory, abdominal and laryngeal adductor muscles. There is further activity by her diaphragm during the compressive phase of the cough reflex that physiologically translates as an isovolume (i.e., no flow) contraction of her chest wall against a closed glottis.

The expulsive phase of the cough reflex begins as the laryngeal adductor muscles (thyroarytenoid and arytenoideus) contract starting a few hundred milliseconds before the diaphragm relaxes and while the abdominal muscles have already relaxed [3, 54, 63]. An effective diaphragmatic force helps my wife to cough. There is inhibition of laryngeal adductor motoneurons, which is important in the generation of explosive expiratory airflow [64]. The posterior cricoarytenoid muscles briefly contract to enlarge the glottic opening but, not as wide as during the inspiratory phase; there is now a strong positive swing of pleural pressure. Her true vocal folds are pulled downwards during the explosive expulsive phase of coughing. The cricothyroid muscle causes vocal fold elongation and increased size of the glottic opening. A transiently relaxed glottis releases a burst of expired air to expel the piece of ice [64]. Finally, the larynx again constricts a bit and her diaphragm relaxes after coughing stops [54, 64].

My wife's true and false cords are participants in coughing since the contraction of the thyroarytenoid muscles alters the position, shape and tension of her false cords. This allows their shelf-like, down-turned free margins to function as a one-way valve to prevent the escape of air from the lower respiratory tract below. In this way, it helps buildup intrathoracic pressure [56]. Her true cords, with their up-turned margins, behave as a one-way valve in the opposite direction, obstructing the entrance of air from above [56]. From a structural point of view, her false cords provide more of an expectorative function for cough, whereas her true cords assume a more protective role against aspiration [56].

My wife's distinctive cough sound is caused by oscillation of the surrounding lung/upper airways' tissues and gases related to the relatively large expiratory airflow velocities during the explosive exhalation [4, 68]. The quality of the cough sound is influenced by airflow speed, changes in resonance of the airway tissues, secretions that are present in the airways and the compliance of the airways [63].

Multitudes of neural messages are integrated, interconnected and funneled to my wife's brainstem cough centers from peripheral afferent sensors that travel through the vagal internal laryngeal nerve to the medulla and interconnect with neural networks in the cortex. Vagal and glossopharyngeal motoneurons innervate her upper airway muscles; the inspiratory and expiratory bulbospinal pre-motoneurons, of the intermediate and caudal regions of the ventral respiratory group, project nerve impulses to phrenic, intercostal, and abdominal motoneurons [4, 58, 69]. Second-order neurons launch signals to her brain stem nervous systems that influence the normal respiratory cycle but also help carry out coughing [58, 70]. The pontine and rostral ventral respiratory groups, the raphe nuclei and Bötzinger and pre-Bötzinger complexes adjust varied cough discharge patterns [58, 70‐72]. Apparently, the pre-Bötzinger complex helps generate inspiratory respiratory rhythm while the retrotrapezoid nucleus/parafacial respiratory group in front of it plays a role for implementing expiratory rhythms. Expiratory neurons, possessing augment firing patterns, span regions of the Bötzinger and pre-Bötzinger complexes to initiate and inhibit premotor neurons of the laryngeal adductor muscles [71, 73‐75]. Possibly, an endogenous cough-suppressing neuronal network located within her caudal ventral respiratory region plays some role in modulating the excitability of the cough reflex [37, 69, 76, 77]; there could be some sort of gating mechanism operating at some level [78, 79].

In humans, the cortical participation of the cough reflex is not like any other animal [5]. Accordingly, my wife is capable of controlling her forced exhalation during coughing. She has the ability to voluntarily initiate or inhibit her cough responses without sensory stimulation, during capsaicin inhalation and also with upper respiratory tract infections [32, 60, 80‐82]. My wife's cough is lost or severely diminished during general anesthesia or sleep; and, systemic opiates suppress her coughing [7, 30, 33, 60, 83]. Her cough is susceptible to placebo-induced suppression [29, 30, 32, 84]. She notes an "urge-to-cough" that always precedes her actual cough motor maneuver [84]. If she is investigated by imaging studies during voluntary coughing, the findings likely will show brain activity appearing in the ventrolateral sensorimotor cortex, an area responsible for non-respiratory orofacial (i.e., chewing, lip pursing and tongue movements) and respiratory orofacial movements (i.e., speaking, singing, and swallowing) [81]. More recently, Mazzone and colleagues, measured blood oxygen level-dependent (BOLD) responses in human subjects utilizing the technique of event-related functional magnetic resonance imaging and confirmed that the largest areas of imaging during voluntary coughing occurred in cortical areas functionally linked to both respiratory-related orofacial tasks (i.e., speaking and singing) and also non-respiratory orofacial actions (i.e., chewing, lip pursing, and tongue movements) [33].