INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

OBSTRUCTIVE SLEEP APNEA is a relatively common clinical condition in which upper airway collapse and cessation of airflow produce an acute cardiovascular response, including elevations in blood pressure as well as changes in heart rate and ventricular function (3, 24, 26). In general, cardiovascular variables fluctuate during the apnea-interapnea periods because of mechanical alterations and reflex changes in autonomic tone. Reflex effects leading to changes in autonomic tone may be activated through 1) changes in thoracic mechanics from obstructed respiratory efforts (4), 2) changes in blood-gas tensions (hypoxia and/or hypercapnia) (12, 18), and 3) postapneic arousals. These factors ultimately culminate in an increase in sympathoadrenal tone (9, 13), which is associated with a pressor response during the apnea and postapnea periods. Studies to date have focused mainly on the relative effect of these physiological alterations with regard to the overall cardiovascular response to apneas. However, few studies have examined specific autonomic pathways mediating this response. In particular, little is known about the relative contribution of vagal traffic in the modulation of the sympathetic response. The vagus nerve could play a prominent role in the cardiovascular response to apneas because of afferent and/or efferent autonomic signals carried within the nerve. Sensory afferent signals from aortic receptors (baroreceptors and/or chemoreceptors) or mechanoreceptors carried via the vagus could be responsible for the augmentation in sympathetic tone during apneas (1, 25). In addition, mechanoreceptor afferents carried via the vagus could increase sympathetically mediated responses through an arousal mechanism (11, 14). On the other hand, parasympathetic efferent signals carried via the vagus nerve could modulate sympathetic activation by decreasing heart rate and lowering blood pressure (3). Thus the vagus could play an important role in mediating the cardiovascular apneic response through both efferent and afferent mechanisms.

In this study, we used a previously developed porcine model of periodic apnea (4) to test the effects of vagal traffic disruption on the acute cardiovascular response to apnea. Previous studies using this model, in which apneas are produced in preinstrumented, sedated, and paralyzed pigs, demonstrated an increase in sympathoadrenal tone as well as a pressor response during the apnea-interapnea cycle (4, 7) compared with preapnea baseline. The model was chosen for this study because of its ability to control for confounding variables such as apnea periodicity, blood-gas and thoracic mechanical alterations during apnea, and acute arousal effects postapnea. We examined the role of vagal traffic in modulating the pressor response associated with apneas by performing bilateral cervical vagotomy. We hypothesized that, because known vagal afferents associated with chemoreceptor (aortic) stimulation are sympathoexcitatory, vagal section would blunt the pressor response to apneas. On the other hand, sympathoinhibatory fibers are also carried by the vagus. If the importance of these were greater than that of sympathoexcitatory fibers, then vagotomy could even have the opposite effect.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study consisted of two phases: a sterile instrumentation phase and a data-collection phase. The local Institutional Animal Care and Use Committee, in accordance with National Institutes of Health guidelines, approved all methods, including protocols, anesthesia, and data collection.

Instrumentation phase. Twelve conditioned female Yorkshire farm pigs, weighing 16-22 kg, were fasted for 12 h before surgery. Surgical plane general anesthesia was established by using ketamine (20 mg/kg) and xylazine (2 mg/kg im). The animals were intubated and mechanically ventilated by using a tidal volume of 12 ml/kg and a respiratory rate of 15 breaths/min. Anesthesia was maintained by administration of halothane (0.5-0.75%) in an enriched oxygen mixture (35-45%). The expiratory port from the anesthesia machine was vented, and all gases were collected by a scavenging system. Heart rate (HR), electrocardiogram (lead II), and oxygen saturation (ear oximeter) were continuously monitored. The animals were placed in the right lateral decubitus position, and 40 mg succinylcholine were administered intramuscularly. By using sterile surgical techniques, an incision was made in the sixth left intercostal space. The fifth rib was cut at or near the sternum to expose the heart and great vessels, and the pericardium was widely excised. After the ascending aorta was carefully separated from the pulmonary artery, a sterile square-wave electromagnetic flow probe (Biotronix) was placed around the ascending aorta (size 14-18 mm, on the basis of aortic diameter). The pericardium was then loosely approximated, and the wire from the flow probe was placed in a sterile plastic bag. After the bag was closed with a rubber band, it was brought through the chest wall and placed in a subcutaneous pocket. A 7-F chest tube was then placed percutaneously, and the chest incision was closed in three layers. The chest was then evacuated by using negative pressure, and the chest tube was removed. The wound was then covered with collodion on sterile gauze, and the anesthesia was discontinued. The animals were allowed to awaken and were extubated once they exhibited strong spontaneous ventilation and chewing movements. The animals were then placed in an individual pen and were allowed to recover. Treatment with pencillin G and benzathine (66,000 U/kg im) during surgery and the following day was administered to prevent wound infection. Postsurgery, morphine sulfate (5 mg im) was given every 6 h during the first 24 h for pain control.

Data-collection phase. Five days postsurgery, the animals were anesthetized with ketamine and xylazine as above. This induced 45-60 min of surgical plane anesthesia, during which time study instrumentation was accomplished and all surgical manipulation was done. In addition, before any incision was made, 2% lidocaine was infiltrated locally. All incisions were closed with silk suture once the appropriate catheters were placed. Animals were intubated and mechanically ventilated by using a tidal volume of 12 ml/kg and a respiratory frequency to maintain arterial PCO₂ in the physiological range. Fluids and medications were administered by using a large-bore catheter placed in the right jugular vein via cutdown. A 7-F catheter was inserted into the femoral artery to measure mean arterial pressure (MAP) and HR and for blood-gas sampling. The subcutaneous pocket was incised, and the aortic flow probe wire was exposed and connected to its recording device. The aortic flow probe was calibrated by using thermodilution cardiac output (CO) measurements from a 7-F pulmonary artery thermistor-tipped catheter placed via the femoral vein. In addition to the recording of continuous CO, the signals from the aortic flow probe were recorded and electronically integrated for beat-by-beat measurement of left ventricular (LV) stroke volume (SV). The vagus nerve was exposed bilaterally in the neck and separated from the sympathetic trunk. Each nerve was tagged for later identification. Once the cutdowns were closed, the animals were placed on the right side.

Continuous sedation. Approximately 20 min before the end of surgical plane anesthesia, a continuous drip (0.9% alphaxalone and 0.3% alphadolone iv; Saffan, Pittman-Moore, Uxbridge, Middlesex, UK) was begun at 3-4 mg · kg¹ · h¹. This level of sedation is sufficient to produce heavy sedation but not surgical plane anesthesia. Previous studies have demonstrated that alphaxalone-alphadolone is associated with preservation of sympathetic reflexes and minimal ventilatory suppression compared with other anesthetic-analgesic agents (2, 10). Once sedation was established, the animals were paralyzed by using cis-atracurium (0.6 mg/kg bolus) followed by a 16 µg · kg¹ · min¹ infusion. We were assured of adequate sedation to eliminate pain or suffering under paralysis in a number of ways. 1) In studies in which animals are not paralyzed (4), but apneas are produced with the same periodicity as in the present protocol (see below), no visible signs of suffering have ever been observed in this laboratory. 2) In previous studies in which apneas were produced in animals both paralyzed and unparalyzed (4), there was no difference in baseline blood pressure or HR, again suggesting that there was no pain or suffering associated with the paralysis and sedation. 3) In the present study, there was continuous monitoring of blood pressure and HR before and after animals were paralyzed, in the presence of continuous sedation. No surges of blood pressure or HR suggestive of pain or suffering were ever observed. 4) Surgical manipulation was only done during surgical plane anesthesia. No surgical manipulation was performed when animals were paralyzed and sedated without surgical plane anesthesia. 5) In previous studies using a protocol similar to that in this study (6), continuous monitoring of electroencephalogram before, during, and after apneas revealed no changes in amplitude or frequency that would indicate arousal or suffering. 6) Heavy sedation was maintained by continuous intravenous drip during all phases of the experiment when paralysis was induced. Animals were euthenized at the end of the experiments by injection 0.3 ml/kg of Euthenasia-5 solution (Veterinary Laboratories, Lenexa, KY), containing pentobarbital sodium (5 g/ml) in 40% isopropyl alcohol and 2% propylene glycol.

Experimental protocol. While the animals were under continuous mechanical ventilation and sedation as described above, periodic apneas were produced by turning off the ventilator for 30 s at end expiration (apnea phase) and then on for 30 s (interapnea phase). Thus a complete apnea-interapnea cycle was defined by 1 min. After an initial baseline period of 20 min, five apnea-interapnea cycles were performed. The animals were then allowed to recover for 15 min. For each condition, measurements were taken at baseline (mechanical ventilation, no apneas), after 5-7 apnea-interapnea cycles had passed from a single apnea-interapnea cycle, and then during recovery (mechanical ventilation, same as baseline conditions). Bilateral cervical vagotomies were then performed on the previously tagged vagus nerves. The above protocol was then repeated. Thus each animal served as its own control.

Data acquisition and calculations. Cardiovascular hemodynamic data recordings were taken at baseline, during the fifth apnea-interapnea cycle, and at recovery both pre- and postvagotomy. Blood-gas data were taken from a sample collected during the fourth or fifth apnea-interapnea cycle 5 s before apnea termination until 5 s into the interapnea interval. Comparison of "baseline" and "recovery" data allowed for assessment of time-related changes. Data variables collected, including HR, MAP, instantaneous CO, and SV, were sampled at a frequency of 200 Hz in 60-s epochs and streamed through to a microcomputer hard disk by using commercially available software (ACQ4600, Gould, Cleveland, OH).

1

5

Statistical analysis. Data were compiled and expressed as means ± SE and percent change ± SE from baseline. Statistical significance between the apnea-interapnea points and recovery period compared with baseline for the control and vagotomy condition was assessed by using a one-way analysis of variance for repeated measures. If significance was found, a post hoc analysis (Newman-Keuls procedure) was used to determine the source of significance. Two-way analysis of variance with planned comparisons was used to test for differences between control and vagotomy for baseline, apnea, and interapnea.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Figure 1 demonstrates the changes in blood-gas tensions for the control and postvagotomy condition. There were no significant differences in PO₂, PCO₂, or pH between baseline and recovery for each condition. As expected, during apneas PO₂ fell significantly, PCO₂ rose significantly, and pH fell significantly compared with baseline in both conditions (P < 0.05). Although there was no significant difference in PO₂ during apnea between conditions, PCO₂ during apnea was slightly but significantly greater in the control condition compared with the vagotomy condition. Accordingly, pH was slightly but significantly less during apnea for control than after vagotomy.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Effects of periodic apneas on pH (A), PO2 (B), and PCO2 (C). Bsln, baseline; Recov, recovery. Values are means ± SE. Changes within any condition (within treatment: control or vagotomy) compared with baseline: * P < 0.05, ** P < 0.01, *** P < 0.001. Changes between control and vagotomy (between treatment) at any given point: # P < 0.05.

Acute hemodynamic effects are shown in Figs. 2-6. For each condition (control, vagotomy) there was no significant difference between baseline and recovery. Thus the stability of the preparation over time was validated. In the control condition, MAP (Fig. 2) increased significantly relative to baseline during apnea and interapnea. However, postvagotomy the MAP response was blunted, with no significant change in MAP occurring relative to baseline. Furthermore, the overall difference (2-way ANOVA) in MAP between control and vagotomy was significant during apnea (P < 0.001), the points of significant difference being at apnea and interapnea. SVR (Fig. 3) behaved similarly to MAP, with a significant (P < 0.05) increase relative to baseline in the control condition during apnea but not postvagotomy. The overall difference between control and vagotomy was significant (2-way ANOVA), the point of significance being during apnea. Figure 4 shows that CO decreased significantly relative to baseline in the control condition during apnea (P < 0.05). However, postvagotomy CO actually increased significantly relative to baseline during apnea and interapnea postvagotomy. The overall difference (2-way ANOVA) between control and vagotomy was also significant (P < 0.01). A significant decrease in SV (P < 0.05) relative to baseline was observed during apnea and interapnea in the control but not the vagotomy condition (Fig. 5). During apnea, SV was significantly lower (P < 0.01) in control vs. vagotomy. Finally, HR (Fig. 6) increased significantly (P < 0.05) relative to baseline during interapnea in the control condition and during both apnea and interapnea postvagotomy. In the control condition the increase in HR was significant between interapnea and apnea (P < 0.05), but after vagotomy the difference in HR between apnea and interapnea was not significant. The difference in HR at apnea was significant between control and vagotomy conditions. Finally, the difference in baseline HR between control and apnea was significant.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Effects of periodic apneas on mean arterial pressure (MAP). Abbreviations and symbols are defined as in Fig. 1.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Effects of periodic apneas on systemic vascular resistance (SVR). Abbreviations and symbols are defined as in Fig. 1.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Effects of periodic apneas on cardiac output (CO). Abbreviations and symbols are defined as in Fig. 1.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Effects of periodic apneas on stroke volume (SV). Abbreviations and symbols are defined as in Fig. 1.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Effects of periodic apneas on heart rate (HR). Abbreviations and symbols are defined as in Fig. 1.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this porcine model of periodic apnea, vagotomy substantially altered the hemodynamic response to apneas during both the apnea and interapnea period. Postvagotomy the pressor response during the apnea-interapnea period was blunted, and there were corresponding modulation in the responses of SVR, SV, and CO. In contrast, the HR response during the apnea and interapnea periods was augmented postvagotomy. The ensuing discussion will consider the study results with respect to the limits of the experimental preparation and presently available literature.

Experimental preparation. A distinguishing feature of obstructive apnea is the persistence of respiratory motoneuron output during the apnea period. In contrast, cessation of respiratory motoneuron output occurs during a central apnea. Thus our porcine model, in which respiratory muscles are paralyzed but respiratory motoneuron output is presumably intact (20), neurologically simulates events occurring during obstructive rather than central apneas and is not to be equated with central apneas. This paralyzed model of periodic apnea was chosen because the pressor response in this model is greater than for a similar nonparalyzed model (4). We hoped, therefore, that perturbation of the system by vagotomy in this model would yield large effects in the variables studied. Because the animals were paralyzed, effects of swings in intrathoracic pressure during apnea on venous return and cardiac function were not present. Thus a potentially confounding factor was controlled for. However, it must be remembered that, in terms of mechano- and proprioceptor sensory neuron feedback, the model resembles central rather than obstructive apneas. The model also controlled for the confounding factor of arousal during apnea. In previous studies, no changes in cortical state suggestive of arousal as measured by electroencephalography were found in this model (6). Finally, it might be thought that baseline hemodynamics with mechanical ventilation would be different from those with spontaneous ventilation and that conclusions about the hemodynamic response to apneas drawn from paralyzed apneas may not apply to baseline conditions for obstructive breathing (spontaneous ventilation). However, Chen and Scharf (4) reported that baseline hemodynamic variables were the same in this model between spontaneous and mechanical ventilation.

Hemodynamic responses during apneas: prevagotomy. The hemodynamic response during the apnea-interapnea period in the control condition included an increase in MAP and SVR, a decrease in CO and SV, and an increase in HR relative to the baseline. The increase in MAP and SVR during apnea and interapnea compared with baseline in the control (prevagotomy) condition is similar to the hemodynamic response seen in humans (8, 21) and observed in previous animal studies (4, 6). We have previously explained the decrease in CO and SV observed during apneas on the basis of increased LV afterload (increased MAP). The present data are consistent with this explanation because the decrease in CO and SV during apneas was in fact seen only when MAP increased.

Hemodynamic responses: effects of vagotomy. Our data clearly demonstrated blunting of the pressor response (MAP) postvagotomy. Instead of decreasing during apnea as it did prevagotomy, CO actually increased during apnea postvagotomy. This increase in CO may have been related, in part, to the effects of apneas on HR.

Mechanism for effects of vagal section on pressor response. Our data demonstrate that the pressor response to apneas in this model is dependent to a large extent on the vagus nerve. Sympathetic activation during periodic apneas is thought to occur from reflex effects elicited from hypoxia, possibly hypercapnia and mechanical alterations (6, 9, 12, 13, 18). These physiological alterations cause activation of peripheral and central sensors, including baroreceptors, chemoreceptors, and mechanoreceptors. Afferent information from these receptors is carried to the central nervous system and ultimately leads to the cardiovascular response. Thus the blunting of the pressor response seen in this study is likely due to loss of the vagal afferent input responsible for eliciting an increase in sympathetic efferent activity. The absence of increased SVR during apneas postvagotomy suggests that vagal nerve traffic prevented sympathetically mediated vasoconstriction. This finding suggests that, at least in pigs, carotid chemo- and baroreceptors may not be as important as aortic receptors in the peripheral sympathetic activation associated with apnea.