Effect of systemic alcohol and nicotine on airway protective reflexes.

OBJECTIVES: Injection of water into the pharynx induces contraction of the upper esophageal sphincter (UES), triggers the pharyngo-UES contractile reflex (PUCR), and at a higher volume, triggers an irrepressible swallow, the reflexive pharyngeal swallow (RPS). These aerodigestive reflexes have been proposed to reduce the risks of aspiration. Alcohol ingestion can predispose to aspiration and previous studies have shown that cigarette smoking can adversely affect these reflexes. It is not known whether this is a local effect of smoking on the pharynx or a systemic effect of nicotine. The aim of this study was to elucidate the effect of systemic alcohol and nicotine on PUCR and RPS. METHODS: Ten healthy non-smoking subjects (8 men, 2 women; mean age: 32+/-3 s.d. years) and 10 healthy chronic smokers (7 men, 3 women; 34+/-8 years) with no history of alcohol abuse were studied. Using previously described techniques, the above reflexes were elicited by rapid and slow water injections into the pharynx, before and after an intravenous injection of 5% alcohol (breath alcohol level of 0.1%), before and after smoking, and before and after a nicotine patch was applied. Blood nicotine levels were measured. RESULTS: During rapid and slow water injections, alcohol significantly increased the threshold volume (ml) to trigger PUCR and RPS (rapid: PUCR: baseline 0.2+/-0.05, alcohol 0.4+/-0.09; P=0.022; RPS: baseline 0.5+/-0.17, alcohol 0.8+/-0.19; P=0.01, slow: PUCR: baseline 0.2+/-0.03, alcohol 0.4+/-0.08; P=0.012; RPS: baseline 3.0+/-0.3, alcohol 4.6+/-0.5; P=0.028). During rapid water injections, acute smoking increased the threshold volume to trigger PUCR and RPS (PUCR: baseline 0.4+/-0.06, smoking 0.67+/-0.09; P=0.03; RPS: baseline 0.7+/-0.03, smoking 1.1+/-0.1; P=0.001). No similar increases were noted after a nicotine patch was applied. CONCLUSIONS: Acute systemic alcohol exposure inhibits the elicitation PUCR and RPS. Unlike cigarette smoking, systemic nicotine does not alter the elicitation of these reflexes.

Manometric evidence for a phonation-induced UES contractile reflex.

The mechanism against entry of gastric content into the pharynx during high-intensity vocalization such as seen among professional singers is not known. We hypothesized that phonation-induced upper esophageal sphincter (UES) contraction enhances the pressure barrier against entry of gastroesophageal contents into pharynx. To determine and compare the effect of phonation on luminal pressures of the esophagus and its sphincters, we studied 17 healthy volunteers (7 male, 10 female) by concurrent high-resolution manometry and voice analysis. We tested high- and low-pitch vowel sounds. Findings were verified in six subjects by UES manometry using a water-perfused sleeve device. Eight of the volunteers (2 male, 6 female) had concurrent video fluoroscopy with high-resolution manometry and voice recording. Fluoroscopic images were analyzed for laryngeal movement. To define the sex-based effect, subgroup analysis was performed. All tested phonation frequencies and intensities induced a significant increase in UES pressure (UESP) compared with prephonation pressure. The magnitude of the UESP increase was significantly higher than that of the distal esophagus, the lower esophageal sphincter (LES), and the stomach. Concurrent videofluoroscopy did not show posterior laryngeal movement during phonation, eliminating a purely mechanical cause for phonation-induced UESP increase. Subgroup analysis demonstrated phonation-induced UESP increases in males that were significantly greater than those of females. Phonation induces a significant increase in UESP, suggesting the existence of a phonation-induced UES contractile reflex. UESP increase due to this reflex is significantly higher than that of the distal esophagus, LES, and stomach. The phonation-induced UESP increase is influenced by sex.

Neurocognitive processing of esophageal central sensitization in the insula and cingulate gyrus.

The cingulate and insular cortices are parts of the limbic system that process and modulate gastrointestinal sensory signals. We hypothesized that sensitization of these two limbic area may operate in esophageal sensitization. Thus the objective of the study was to elucidate the neurocognitive processing in the cingulate and insular cortices to mechanical stimulation of the proximal esophagus following infusion of acid or phosphate buffer solution (PBS) into the esophagus. Twenty-six studies (14 to acid and 12 to PBS infusion) were performed in 20 healthy subjects (18-35 yr) using high-resolution (2.5 x 2.5 x 2.5 mm(3) voxel size) functional MRI (fMRI). Paradigm-driven, 2-min fMRI scans were performed during randomly timed 15-s intervals of proximal esophageal barostatically controlled distentions and rest, before and after 30-min of distal esophageal acid or PBS perfusion (0.1 N HCl or 0.1 M PBS at 1 ml/min). Following distal esophageal acid infusion, at subliminal and liminal levels of proximal esophageal distentions, the number of activated voxels in both cingulate and insular cortices showed a significant increase compared with before acid infusion (P < 0.05). No statistically significant change in cortical activity was noted following PBS infusion. We conclude that 1) acid stimulation of the esophagus results in sensitization of the cingulate and insular cortices to subliminal and liminal nonpainful mechanical stimulations, and 2) these findings can have ramifications with regard to the mechanisms of some esophageal symptoms attributed to reflux disease.

Safety and feasibility of evaluating airway-protective reflexes during sleep: new technique and preliminary results.

BACKGROUND: The airway is vulnerable to aspiration during sleep. The integrity of aerodigestive-protective reflexes during sleep has not been studied previously because of a lack of adequate techniques. OBJECTIVE: To determine the safety and the feasibility of a new technique to elicit pharyngoglottal closure reflex (PGCR), pharyngo-upper-esophageal sphincter (UES) pressure contractile reflex (PUCR), and reflexive pharyngeal swallow (RPS) during sleep. SETTING: Outpatient sleep laboratory. DESIGN AND INTERVENTION: PGCR, PUCR, and RPS were elicited in 3 subjects by injecting colored water into the pharynx through a specially designed UES manometry catheter to which a thin videoendoscope was taped. This assembly was passed transnasally and positioned to obtain UES-pressure recordings and adequate endoscopic glottic views. Sleep was monitored by polysomnography, and all modalities were synchronized by using a timer. Subjects were evaluated while awake and during stage I sleep. RESULTS: All subjects were monitored for 3 hours of natural sleep, during which several periods of stage I sleep were observed. While awake, PGCR, PUCR, and RPS were elicited in all subjects. During sleep, PGCR was present in all, PUCR in 2, and RPS in 2 (1 after arousal) subjects. Threshold volumes for reflex elicitation were not significantly different between the awake state and stage I sleep. None of the subjects exhibited laryngeal penetration or aspiration. LIMITATIONS: Small numbers of subjects were studied only in stage I sleep. CONCLUSIONS: When using the above technique, it is safe and feasible to study aerodigestive reflexes during sleep. Preliminary data suggest that PGCR, PUCR, and RPS can be elicited during sleep.

Performance and optimal technique for pharyngeal impedance recording: A simulated pharyngeal reflux study.

BACKGROUND AND AIMS: Detection rate, influence of recording site, and subject posture for impedance monitoring of pharyngeal reflux of gastric contents remain unknown. We evaluated the ability of the impedance sensor for detection of various volumes of intrapharyngeal infusate at two sites and in two subject positions. METHODS: Nineteen healthy subjects were studied using concurrent videoendoscopic, manometric, impedance, and pH recording. RESULTS: Detection rate of simulated pharyngeal reflux events ranged between 87% and 100% for 1-4 mL. Detection rate for 0.1-1 mL volumes in the upright position was significantly higher (78-85%) when the impedance sensor was located at the proximal margin of the upper esophageal sphincter (UES) compared to 2 cm proximally (38-68%) (P < 0.001). With the sensor at 2 cm above the UES, the average detection rate for all volumes in the upright position was significantly less (P < 0.001) compared to the supine position (48%vs 84%). There was substantial variability in the magnitude of impedance changes induced by different infusates. CONCLUSIONS: Impedance sensors can detect as small a volume as 0.1 mL and combined with a pH sensor can detect acidic and nonacidic liquid and mist reflux events. Sensor placement at the proximal margin of the UES yields the highest detection rate irrespective of subject posture compared to placement 2 cm proximally. Depending on the volume of refluxate and location of the impedance sensor, a substantial minority of simulated reflux events can be missed.

Impedance signature of pharyngeal gaseous reflux.

BACKGROUND: Pharyngeal impedance changes induced by various pharyngeal reflux events have not been characterized. OBJECTIVES: To characterize pharyngeal impedance changes induced by participant-perceived belching events. METHODS: We systematically evaluated pharyngeal impedance and pH changes related to 453 belch events in 11 gastroesophageal reflux disease, 10 reflux attributed-laryngitis patients and 16 controls. RESULTS: Of 453 belch events, 362 were analyzable. Of these, 72% occurred within 10 s, 93% within 20 s, 99% within 30 s and 100% within 40 s of the time that participants marked a belch event. In 15% impedance changes in the pharynx preceded, in 12% they were simultaneous and in 73% they occurred after the start of the impedance change in the proximal esophagus. Time interval between the two events ranged between 0.4+/-0.03 and 0.7+/-0.1 s. In all, there were three types of belch-induced impedance changes: (a) impedance increase, (b) impedance decrease and (c) multiphasic. Twenty percent of impedance events associated with belching had less than 50% change from baseline, whereas in 51% changes exceeded or were equal to 50%. Among events with a drop in pharyngeal impedance, only two satisfied the criteria for the liquid reflux event. CONCLUSIONS: Pharyngeal ventilation of gastric gaseous content seems to have a unique impedance signature. During pharyngeal gas reflux events, impedance changes may start before or after proximal esophageal changes. Belching may induce negative pharyngeal changes that do not meet the criteria for liquid reflux. These findings need to be taken into consideration in the analysis of pharyngeal reflux events.

Effect of aging on bolus kinematics during the pharyngeal phase of swallowing.

Swallowing difficulty is a common complaint in the elderly and, although there are data for the biomechanics of liquid swallows, little is known about solid bolus motion, or kinematics, in the elderly. The aims of this study were as follows: 1) to characterize and compare solid and liquid bolus kinematics in the elderly and compare the findings with those in young subjects and 2) to correlate bolus kinematics and dynamics. Concurrent manometric-fluoroscopic techniques were used to study eight young and eight elderly subjects. The subjects performed four swallows each of 0.2-cm-diameter solid barium pellets and 5 ml of liquid barium during sagittal fluoroscopy and six-channel pharyngoesophageal manometry. Images were digitized for analysis of kinematic properties such as velocity and acceleration. Dynamic pressures were recorded and coordinated with kinematic events. Image analysis showed that velocity varied as the pellet passed through the hypopharynx, pharynx, and upper esophageal sphincter. In young subjects, pellet kinematics were characterized by two zones of pellet acceleration: one over the tongue base and another as the pellet passed through the upper esophageal sphincter. Although the elderly showed a similar zone of acceleration over the base of the tongue, the second zone of pellet acceleration was not seen. Decreasing pressure gradients immediately distal to the position of the solid pellet and liquid bolus characterized dynamics for all subjects. This decreasing pressure gradient was significantly larger in elderly than in young subjects. Bolus kinematics and dynamics were significantly altered among elderly compared with young subjects. Among these differences were the absence of hypopharyngeal bolus acceleration and a significant increase in the trans-sphincteric pressure gradient in the elderly.

Influence of sleep stages on esophago-upper esophageal sphincter contractile reflex and secondary esophageal peristalsis.

BACKGROUND & AIMS: Airways are most vulnerable to aspiration during sleep. Esophago-upper esophageal sphincter (UES) contractile reflex (EUCR) and secondary peristalsis (2P) have been proposed to protect the airway by reflexively contracting the UES and clearing the esophagus of refluxate, respectively. Our aim was to study EUCR and 2P elicitation in "awake" state, stage II, slow-wave (stage III/IV), and rapid eye movement (REM) sleep. METHODS: Thirteen healthy volunteers were studied in the supine position using concurrent UES and esophageal manometry and polysomnography. Threshold volume (Tvol) to trigger EUCR and 2P and changes in sleep stages were recorded during injection of 2.7 mL/min water into the proximal esophagus after sleep stages were confirmed. RESULTS: UES pressure progressively declined with deeper stages of sleep. Tvol for EUCR and 2P elicitation was not significantly different between the stage II and "awake" state (EUCR: 4.0 +/- 1.8 mL vs 6.1 +/- 3.6 mL stage II; 2P: 5.8 +/- 2.2 mL vs 8.0 +/- 4.0 mL stage II). Tvol for EUCR and 2P elicitation during REM sleep were significantly lower than during the stage II and "awake" state (REM EUCR: 2.2 +/- 1.1 mL; 2P: 3.5 +/- 1.2 mL). Arousal and cough preempted development of EUCR and 2P during slow-wave sleep. CONCLUSIONS: (1) EUCR/2P can be elicited in stage II and REM but is preempted by arousal in slow-wave sleep. (2) Tvol for EUCR/2P elicitation is significantly lower in REM, compared with the stage II and "awake" state, suggesting a heightened sensitivity of these reflexes during REM sleep. (3) Although UES pressure progressively declines with deeper stages of sleep, it can still reflexively contract during REM sleep, despite generalized hypotonia.

Novel evidence for hypersensitivity of visceral sensory neural circuitry in irritable bowel syndrome patients.

BACKGROUND & AIMS: Visceral hypersensitivity in irritable bowel syndrome (IBS) patients has been documented by evaluation of perceived stimulations that can reflect abnormalities of both sensory neurocircuitry and cognitive processes. The presence of actual neurohypersensitivity in human beings has not been documented separately. Because subliminal stimulations are free from the influence of stimulus-related cognitive processes, functional magnetic resonance imaging (fMRI) cortical response to these stimuli can be considered a measure of activity of the neural circuitry alone. The aim of this study was to compare quantitatively the cerebral cortical fMRI activity response to equal subliminal stimulations between IBS patients and age-matched controls. METHODS: We studied 10 IBS patients and 10 healthy controls using a computerized barostat-controlled rectal distention device. fMRI activity volume and percent maximum signal intensity change for equal subliminal distention pressures were compared between controls and patients. RESULTS: Three levels of subliminal distention pressures (eg, 10, 15, and 20 mm Hg), were represented in both controls and patients and were analyzed for fMRI response. In all 3 distention levels the fMRI activity volume in IBS patients was significantly larger than age- and sex-matched controls (P < .05). The percent maximum signal intensity change was similar between IBS patients and controls. CONCLUSIONS: The volume of cerebral cortical activity response to equal subliminal distention pressures in IBS patients is significantly larger than in controls, documenting the existence of hypersensitivity of the neural circuitry in this patient group irrespective of stimulus-related cognitive processes.

Laparoscopic Nissen fundoplication decreases gastroesophageal junction distensibility in patients with gastroesophageal reflux disease.

Laparoscopic Nissen fundoplication (LNF) is the surgical treatment of choice for gastroesophageal reflux disease (GERD). Post-LNF complications, such as gas bloat syndrome, inability to belch and vomit, and dysphagia, remain too common and prevent LNF from being more highly recommended. It remains controversial as to whether preoperative assessment can predict the development of post-LNF complications. Some authors have shown a correlation between pre-LNF manometry characteristics and post-LNF dysphagia, and others have not. We hypothesize that many post-LNF complications are caused by a decrease in the distensibility of the GEJ and that standard manometry is at best an indirect measure of this. The aim of this study is to directly measure the effect of LNF on gastroesophageal junction (GEJ) distensibility (GEJD). The lower esophageal sphincter (LES) of 15 patients undergoing LNF was characterized using standard manometry. The GEJD before and after a standardized LNF was measured using a specialized catheter, containing an infinitely compliant bag, placed within the LES. GEJD was measured, as dV/dP over volumes 5 to 25 mL distended at a rate of 20 mL/min. Mean dP +/- standard error of the mean for each volume was calculated, and distensibility curves were generated and compared. Measurements were also taken after abolishing LES tone by mid-esophageal balloon distension. Patient symptoms were recorded before and after surgery. Statistical analysis was performed by two-way repeated-measures analysis of variance, paired t test, and the Tukey test. Laparoscopic Nissen fundoplication led to a statistically significant increase in Delta pressure over each volume tested and therefore a significant decrease in the distensibility of the GEJ. Abolition of LES tone had no statistical effect on GEJD after fundoplication. There were no complications, and none of the patients developed the symptom of dysphagia postoperatively. These are the first direct measurements to show that LNF significantly reduces the distensibility of the GEJ. We hypothesize that the magnitude of this reduction may be the vital variable in the development of post-LNF complications and specifically post-LNF dysphagia. The intraoperative measurement of LES distensibility may provide a means for avoiding this feared and other post-LNF complications in the future.

Cingulate cortex: a closer look at its gut-related functional topography.

Earlier studies have documented activation of the cingulate cortex during gut related sensory-motor function. However, topography of the cingulate cortex in relationship to various levels of visceromotor sensory stimuli and gender is not completely elucidated. The aim was to characterize and compare the activation topography of the cingulate cortex in response to 1) subliminal, 2) perceived rectal distensions, and 3) external anal sphincter contraction (EASC) in males and females. We studied 18 healthy volunteers (ages 18-35 yr; 10 women, 8 men) using functional MRI blood-oxygenation-level-dependent technique. We obtained 11 axial slices (voxel vol. 2.5-6.0 x 2.5 x 2.5 mm(3)) through the cingulate cortex during barostat-controlled subliminal, liminal, and supraliminal nonpainful rectal distensions as well as EASC. Overall, for viscerosensation, the anterior cingulate cortex exhibited significantly more numbers of activated cortical voxels for all levels of stimulations compared with the posterior cingulate cortex (P < 0.05). In contrast, during EASC, activity in the posterior cingulate was larger than in the anterior cingulate cortex (P < 0.05). Cingulate activation was similar during EASC in males and females (P = 0.58), whereas there was a gender difference in anterior cingulate activation during liminal and supraliminal stimulations (P < 0.05). In females, viscerosensory cortical activity response was stimulus-intensity dependent. Intestinal viscerosensation and EASC induce different patterns of cingulate cortical activation. There may be gender differences in cingulate cortical activation during viscerosensation. In contrast to male subjects, females exhibit increased activity in response to liminal nonpainful stimulation compared with subliminal stimulation suggesting differences in cognition-related recruitment.

Optimal stimulus intensity and reliability of air stimulation technique for elicitation of laryngo-upper esophageal sphincter contractile reflex.

To determine the optimal air stimulus intensity and duration for elicitation of the laryngo-upper esophageal sphincter (UES) contractile reflex, we studied 37 healthy volunteers 20 to 81 years of age. A sleeve device monitored the UES pressure. For laryngeal stimulation, we used an air stimulator unit (Pentax AP-4000) that incorporated a nasolaryngeal endoscope. The arytenoids and interarytenoid areas were stimulated at least three times by three different stimuli: 6-mm Hg air pulse with 50-ms duration, 10-mm Hg air pulse with 50-ms duration, and 6-mm Hg air pulse with 2-second duration. Of 1,165 air stimulations, 1,041 resulted in mucosal deflections. Of these, 451 resulted in an abrupt increase in UES pressure. The response/deflection ratio for 6-mm Hg stimulation with 2-second duration was significantly higher than those for air pulses with 50-ms duration (p < .001). We conclude that although the laryngo-UES contractile reflex can be elicited by an air pulse with 50-ms duration, this ultrashort stimulation is not reliable. Using longer-duration pulses (at least 2 seconds) improves the reliability of elicitation of the laryngo-UES contractile reflex.

Laryngo-upper esophageal sphincter contractile reflex in humans deteriorates with age.

BACKGROUND & AIMS: Recent studies have shown the existence of several reflex connections between the aerodigestive and upper gastrointestinal tracts. Our aim was to study the effect of laryngeal stimulation on upper esophageal sphincter (UES) pressure and to determine the reproducibility of this effect. METHODS: We studied 14 young and 10 elderly healthy nonsmoker volunteers and 7 patients with UES dysphagia using a concurrent manometric and video endoscopic technique. Three levels of laryngeal air stimulation were studied: 6 mm Hg/50 ms, 10 mm Hg/50 ms, and 6 mm Hg/2 s. Ten young subjects were studied twice. RESULTS: For 6-mm Hg/2-s and 6-mm Hg/50-ms duration stimuli, the frequency of UES response to air stimulation as evidenced by mucosal deflection (response/deflection ratio) in the elderly volunteers was significantly lower compared with that of young subjects (P < 0.05). The response/deflection ratio of the 6-mm Hg/2-s stimulus was significantly higher than those induced by stimuli of shorter duration (P < 0.01). Poststimulation UES pressure was significantly higher than prestimulation pressure (P < 0.05) in both groups. The magnitude of the increase in poststimulation UES pressure in the elderly volunteers was similar to that of the young subjects. Findings were similar in repeated studies. Four of 7 dysphagic patients exhibited an abnormal response. CONCLUSIONS: Afferent signals originating from the larynx reproducibly induce contraction of the UES: the laryngo-UES contractile reflex. This reflex is elicited most reliably by 6-mm Hg/2-s air stimulation. Frequency elicitation of this reflex decreases significantly with age while the magnitude of change in UES pressure remains unchanged, indicating a deleterious effect of aging on the afferent arm of this reflex. This reflex is altered in some dysphagic patients.

Physical and pH properties of gastroesophagopharyngeal refluxate: a 24-hour simultaneous ambulatory impedance and pH monitoring study.

OBJECTIVES: Frequency occurrence of nonacidic and nonliquid reflux events in the pharynx has not been systematically studied. The aim of the present study was to characterize the physical (liquid, gas, and mixed gas/liquid) and pH properties of the gastroesophagopharyngeal refluxate. METHODS: We performed a total of 31 24-h simultaneous ambulatory pharyngoesophageal impedance and pH recordings in 11 GERD patients, 10 patients with reflux-attributed laryngitis, and 10 healthy controls. RESULTS: On average, the total number of reflux events (all kinds) in the pharynx was less than half of that in the proximal esophagus (18 +/- 4 vs 50 +/- 4, p < 0.01). Most of the pharyngeal reflux events were gas events and were observed in all three studied groups. Prevalence of these gas reflux events ranged between 0 and 74. The number of gas reflux events accompanied by a minor pH drop in laryngitis patients (1 (0-36)) was significantly higher than those in GERD and controls (0 (0-2) and 0 (0-1), respectively, p < 0.05). There was no significant difference in the number of nonacidic gas reflux events among the three groups (GERD: 10 (2-57), laryngitis: 11.5 (0-51), controls: 10.5 (0-27)). Impedance recording identified a total number of 566 events in the pharynx. Of these, a total of 563 events were compatible with gas reflux events, 101 events were accompanied by minor drops in intrapharyngeal pH, whereas 460 events were not accompanied by any pharyngeal pH change. CONCLUSIONS: Concurrent impedance and pH recordings detect significantly more events qualifying as reflux in the pharynx than pH recordings alone. A substantial majority of these events are gaseous refluxes both with and without minor pH drops. Gas reflux events with weak acidity appear to be more common among patients with reflux-attributed laryngeal lesions compared to GERD patients and controls.

Characterization of the cerebral cortical representation of heartburn in GERD patients.

Although symptoms arising from the esophagus such as heartburn and pain can at times become challenging clinical problems, esophageal viscerosensation, especially with regard to chemical stimulation in humans, is incompletely understood. Our aims were 1) to characterize and ascertain the reproducibility of cerebral cortical registration of heartburn and 2) to elucidate the differences between these findings and those of esophageal subliminal acid stimulation in asymptomatic controls. We studied 11 gastroesophageal reflux disease (GERD) patients (9 males, 30-55 yr) and 15 healthy controls (8 males, 21-49 yr). Cerebral cortical functional magnetic resonance imaging (fMRI) activity was recorded twice in each subject, during two 5-min intervals of 0.1 N HCl, separated by 5 min of NaCl perfusion. Time from onset of acid perfusion to instant of fMRI signal increase and first report of heartburn averaged 1.60 +/- 0.80 and 1.85 +/- 0.60 min, respectively. Average maximum percent signal increase in the GERD patients (16.3 +/- 3.5%) was significantly greater than that of healthy controls (3.8 +/- 0.9%; P < 0.01). Temporal fMRI signal characteristics during heartburn were significantly different from those of subliminal acid stimulation in controls (P < 0.01). Activated cortical regions included sensory/motor, parieto-occipital, cingulate and prefrontal regions, and the insula. There was 92% concordance between the activated Brodmann areas in repeated studies of GERD patients. Cortical activity associated with perceived and unperceived esophageal acid exposure in GERD patients and healthy controls, respectively, involves multiple brain regions but occurs more rapidly and with greater intensity in GERD patients than the activity in response to subliminal acid exposure in healthy controls. The cortical pain-processing pathway seems to be involved in perception of esophageal acid exposure and could explain the variations encountered in clinical practice defining this sensation.