A Comparison of Baroreflex Sensitivity during Isoflurane and Desflurane Anesthesia in Humans

After Human Research Review Committee approval, 18 healthy, normotensive, male volunteers (aged 20-31 yr) provided informed consent and were instrumented with electrocardiogram, peripheral intravenous and radial arterial catheters, and microneurographic needle electrodes, as described elsewhere. [20]Subjects fasted for at least 12 h before the study but received 30 ml of oral sodium bicitrate as a precautionary measure. Subjects received 5-7 ml *symbol* kg sup -1 of intravenous saline before measurements. Basal HR, BP, and efferent SNA directed to blood vessels within skeletal muscle were recorded at rest after a 20-min accommodation period following instrumentation.

Electrocardiogram signals were processed by computerized identification of R spikes and measurement of RR intervals. Efferent nerve signals were amplified (x100,000), band-pass filtered (200-2,000 Hz), rectified, and integrated (150 ms moving averager) to produce a processed nerve signal. Individual "bursts" of integrated nerve activity were identified by computer and confirmed manually over 5-min epochs at each steady-state anesthetic level. Sympathetic nerve activity was identified by previously defined criteria. [5,20]This activity has been shown to be similar in upper and lower, ipsilateral and contralateral extremities. [20]Because muscle blood flow is 40% of resting cardiac output, recordings of muscle SNA represent an important window on neural control of BP. SNA often is quantified as the number of bursts per minute. A better method to express SNA is by quantitating the number of integrated bursts that occur every 100 heartbeats, because the integrated bursts of SNA are pulse-synchronous with the electrocardiogram, and HR varies widely over the stress period. Integrated bursts also vary in amplitude based on the recruitment of sympathetic fibers each heartbeat. Thus, we also define muscle SNA as total activity, which is the average burst amplitude (micro Volt) measured over the sampling period multiplied by the number of burst per 100 heartbeats. Because the amplitude of sympathetic bursts can be altered by unpredictable/unwanted needle movement within the nerve, both methods of quantification are reported.

The activity from the baroreceptors was altered via drug-induced, sequential hypo- and hypertensive challenges. A bolus injection of sodium nitroprusside (100 micro gram) was followed 60 seconds later by a bolus injection of phenylephrine (150 micro gram). These doses were designed to produce a range of BP values within physiologically acceptable limits. The doses chosen caused pressure to be decreased by about 15 mm Hg and then increased approximately 10 mm Hg above control levels. Sodium nitroprusside and phenylephrine were employed because of their direct acting effects with minimum direct influence on the reflex pathways under study.

After recording neurocirculatory parameters at conscious baseline and the response to the baroreceptor challenge, subjects breathed oxygen, and anesthesia was induced with 100 mg thiopental (to reduce myoclonic responses), followed immediately by 0.3 mg *symbol* kg sup -1 etomidate and 0.2-0.25 mg *symbol* kg sup -1 vecuronium to facilitate intubation. After tracheal intubation, the anesthetic was randomly chosen to be either isoflurane (n = 9) or desflurane (n = 9). Anesthesia was maintained with isoflurane or desflurane at 0.5 (0.6 or 3.6%), 1.0 (1.2 or 7.2%), or 1.5 (1.8 or 10.9%) MAC in a randomized fashion. End-tidal anesthetic and carbon dioxide levels were continuously monitored (infrared spectrometry, Ohmeda, WI). Ventilation was adjusted to maintain P^CO2at 38-40 mm Hg, which was confirmed at each anesthetic level by an arterial blood gas measurement. After 30 min of equilibration to allow the effects of the induction agents to subside, steady-state and baroreceptor stress measurements were repeated at the first MAC setting. The anesthetic depth was adjusted to the next randomized level, and another 20-min equilibration period was allowed before additional measurements were repeated. Flow rates were adjusted to achieve the target end-tidal anesthetic concentration within the first 10 min of the 20-min period.

Cardiac baroslopes were derived by plotting RR interval versus mean arterial pressure (MAP) over the stress period induced by nitroprusside and phenylephrine (Figure 1). To more easily manage the large number of points and nonbaroreceptor-mediated variations of RR interval (due to respiration and arousal), RR interval data were averaged for each sequential 2 mm Hg of BP before plotting RR interval versus MAP, as shown in Figure 1. The linear portions of the typically sigmoid relation were derived by computer after visual selection of the general region of interest. The HR response to increasing BP did not typically overlay the response to decreasing BP, an effect known as hysteresis. (This may be due to a differential baroresponse based on the direction of pressure change.) Therefore, the linear portions of the response curve were fitted to separate lines by regression analysis for both falling and rising pressure responses, as shown in Figure 1. Similar sympathetic baroslopes were created from the SNA response to lowered diastolic pressure. The sympathetic response represents primarily the falling pressure response, because as pressure was elevated above baseline, nerve activity was shut off. The choice of MAP and diastolic pressure as the ordinate in the regression analyses was based on published data demonstrating that these variables are the primary determinants of arterial baroreflex neuroeffector responses in healthy humans. [21].

Consecutive hemodynamic and neural measurements were compared by analysis of variance for repeated measures, and paired comparisons were made with Dunnett's t tests. Differences between baroslopes derived under desflurane and isoflurane anesthesia were compared by analysis of variance for repeated measures. Probability values less than 0.05 were considered sufficient to reject the null hypothesis for all comparisons.

The hemodynamic measurements at conscious baseline and during steady-state levels of anesthesia are displayed in Table 1and are consistent with what has been described previously by this laboratory in this population. [2]In two subjects receiving isoflurane at 1.5 MAC, MAP decreased below 50 mm Hg, and baroreflex testing was not performed. Their data were dropped from the final study analysis and are not included in the presentation. The awake, resting HR, MAP, and SNA were similar between groups. Increasing MAC of isoflurane resulted in significant increases in HR from conscious baseline. Conversely, 0.5 and 1.0 MAC desflurane did not result in changes in HR, but 1.5 MAC desflurane was associated with significant increases in HR similar to the isoflurane group (Table 1). Resting MAP was significantly reduced in a dose-dependent fashion by both anesthetics, and there were no differences between anesthetics, SNA initially decreased at 0.5 MAC desflurane and gradually increased at higher MAC. Although SNA at 0.5 MAC isoflurane also decreased from conscious baseline, it did not significantly increase with higher MAC, and at 1.5 MAC, it was significantly lower than equipotent desflurane.

Awake cardiac and sympathetic baroslopes were similar between groups (Table 2). In all cases, cardiac baroslopes derived during decreasing pressure were less than responses to increasing pressure. Isoflurane and desflurane produced similar decreases in cardiac baroslopes in response to the increasing and decreasing pressure stimuli (Figure 2). The sympathetic slopes differed between anesthetics only at the 0.5 MAC level, with isoflurane showing a preserved slope and desflurane a decreased response (Table 2and Figure 3). Baroslopes decreased in magnitude with increasing anesthetic depth except for isoflurane at the 0.5 MAC level, as mentioned above.

In this study, we assessed the ability of the baroreceptor reflex to respond to hypo- or hypertensive stresses during increasing MAC levels of desflurane and isoflurane anesthesia. From previous work, [7-9,11,14,19]we learned that most volatile anesthetics suppress the ability to reflexively adjust BP and HR in response to BP perturbations. Seagard et al., [10]in their work with dogs, reported that volatile anesthetics act at several sites along the baroreceptor pathway to produce these results, including the baroreceptors, afferent and efferent nerve pathways, central nervous system, peripheral ganglia, and heart. Although most components of the baroreflex were suppressed by volatile anesthetics, they found that baroreceptor afferent activity was increased with deeper anesthetic levels. [10]The increased inhibitory afferent activity was thought to partially account for the decrease in BP observed with these agents.

Several observations suggest that baroreflex control of the circulation may differ during desflurane anesthesia compared to isoflurane. Although preliminary work by Stekiel et al. [22]suggests that the reflex mesenteric vascular response of rabbit to baroreceptor stimulation or inhibition is similarly attenuated by desflurane, halothane, and isoflurane, observations in humans suggest that isoflurane and desflurane might have dissimilar effects on the reflex control of the circulation. We [2]and others [1-3]. have shown that equi-MAC concentrations of isoflurane and desflurane produce similar decreases in BP. However, isoflurane produces a moderate increase in HR at 0.5 MAC, whereas desflurane does not change HR. Moreover, despite similar BP decreases at 1.5 MAC, desflurane is associated with a greater level of SNA and greater plasma norepinephrine concentrations than those seen with isoflurane. [2,3]One possible explanation for these differences has been that isoflurane preserves baroreflex function at low MAC concentrations, allowing for reflex increases in HR in response to hypotension, whereas at 1.5 MAC desflurane, the inhibitory afferent nerve traffic from baroreceptors is impaired, leading to augmented sympathetic outflow.

In the current study, the cardiac baroreceptor reflex was decreased similarly and in graded fashion by both anesthetic agents, suggesting a reduced ability to increase HR for any given decrease in pressure. Despite similar decreases in BP, isoflurane was associated with an increase in HR at 0.5 MAC, and desflurane was not. This effect cannot be attributed to differences in cardiac baroreflex responsiveness. However, the sympathetic baroslopes were preserved at 0.5 MAC isoflurane but decreased at similar levels of desflurane. This finding might explain the early tachycardia in response to isoflurane at 0.5 MAC but would not explain the disparate SNA responses at deeper anesthetic levels. If the sympathetic reflex activity were responsible for the increased HR response, one would expect a similar magnitude of HR increase for both agents at the 1.0 MAC level when baroreflex function was diminished similarly. The inability to directly relate changes in sympathetic baroslopes with alterations in basal HR at different anesthetic depths is related in part to human baroreflex physiology: The regulation of HR is almost entirely via activation or withdrawal of parasympathetic activity to the SA node. [23,24]Thus, the differences in steady-state HR response produced by these two agents cannot be explained by differences in sympathetic baroreflex responsiveness.

A second reason that baroreflex activity may not correlate with basal levels of HR and BP can be found in the structure of the baroreceptor HR control system. The observed differences in HR at various depths of anesthesia were made at steady-state, but the reflex measurement technique was in response to a stress on the system at each steady-state anesthetic concentration. The dynamic testing procedures that we employed may not appropriately assess the autonomic influences on basal HR or SNA under anesthesia but instead reflect the ability to oppose acute pressure changes at these new set-points. Such a distinction between basal autonomic activity versus rapid reflex adjustments in autonomic outflow has been strengthened by recent findings of Seagard and colleagues [25-26]of two distinct sets of carotid baroreceptor afferent cell types. Type I cells are generally quiescent but respond briskly to dynamic pressure changes and can quickly reset to operate at new pressure thresholds. Type II cells are tonically active, fire in proportion to the absolute pressure, and slowly reset. Whether type I or II cells in the carotid baroreceptors have differential effects on efferent vagal or SNA responses has not been determined; however, we conjecture that isoflurane and desflurane might similarly alter the activity of dynamic type I cells to depress the acute reflex response to pressure changes with increasing depth of anesthesia. Furthermore, the slower reacting, tonic limb of the baroreflex arc (perhaps mediated by type II cells) may be dissimilarly altered by desflurane and isoflurane, leading to different basal HR and SNA with increasing anesthetic concentrations.

Clinically, our data confirm the ability of volatile anesthetics to progressively suppress the acute reflex responses to hypo- or hypertensive stimuli with increasing anesthetic depth, and the effect appears to be greater for decreasing versus increasing pressure responses. In addition to attenuating the neural reflexes, isoflurane and desflurane lower BP by direct effects on vascular smooth muscle. This would alter vascular dispensability and deformation of the vessel wall containing the baroreceptor sensors, as demonstrated previously in dogs, where controlled pressure stimuli resulted in enhanced afferent nerve activity from baroreceptors during isoflurane administration compared to control conditions. [10]In the human model, it is virtually impossible to evaluate the direct effects of isoflurane and desflurane on the vascular wall containing baroreceptors and on the afferent profile from these sites during pressure loading and unloading conditions.

In summary, we found a graded depression of both pressor and depressor cardiac baroslopes with increasing anesthetic depth, and there were essentially no differences between isoflurane and desflurane. There was a similar depression of sympathetic baroslopes with increasing MAC, with the exception that low-dose isoflurane maintained sympathetic reflexes near awake levels. These findings do not explain the different steady-state HR and SNA responses to isoflurane and desflurane anesthesia.

*Eckberg DI., Fritsch JM: How should human baroreflexes be tested? News in Physiological Sciences 8:7-12, 1993.