Influence of baroreflex-mediated tachycardia on the regulation of dynamic cerebral perfusion during acute hypotension in humans

Abstract

The effect of acute arterial baroreflex dysfunction on cerebral autoregulation (CA) in otherwise healthy humans is unknown. We identified dynamic CA with and without arterial baroreflex-mediated tachycardia and consequent changes in cardiac output during acute hypotension whilst continuously monitoring changes in middle cerebral artery mean blood velocity (MCA V_mean). Acute hypotension was induced in nine healthy subjects (mean ±s.d.; 26 ± 3 years) by releasing bilateral thigh cuffs after 6 min of supra-systolic resting ischaemia. Hypotension was induced before and after sympathetic blockade (β-1 receptors), and combined sympathetic–cholinergic blockade. That sequential bolus injections of sodium nitroprusside (50 μg), followed 60 s later by phenylephrine hydrochloride (50 μg), elicited < 5 beats min⁻¹ change in heart rate was verified to confirm that full cardiac autonomic blockade was achieved. Thigh cuff release elicited a transient drop in mean arterial pressure and resultant tachycardia. This tachycardic response was diminished in full cardiac blockade (vs. control, P= 0.029; vs.β-1 adrenergic blockade, P= 0.031). Dynamic CA was also attenuated in the full blockade condition compared to both control (P= 0.028) and β-1 adrenergic blockade conditions (P= 0.015), and was related with the attenuated tachycardia response (P= 0.015). These data indicate an important role of the cardiac baroreflex in dynamic CA.

Introduction

Cerebral blood flow (CBF) is traditionally thought to remain relatively constant within a large range of pressures (60–150 mmHg) – static cerebral autoregulation (CA; Lassen, 1959). By releasing bilateral thigh cuffs to elicit transient hypotension, the simultaneous recording of middle cerebral artery blood velocity (MCA V) using transcranial Doppler (TCD) demonstrated that regulation of CBF involved a dynamic component distinct from static CA (Aaslid et al. 1989). It was proposed that ‘dynamic CA’ could be quantified using the rate of regulation (RoR) index, a metric based on the rate of change in cerebral vascular resistance taking place within a 2.5 s window, 1 s following cuff release.

The thigh cuff release has been applied for the evaluation of dynamic CA because it is non-invasive and non-pharmacological (Tiecks et al. 1995; Ogoh et al. 2008). However, thigh cuff release elicits an integrated response that involves not only cerebrovascular changes but also arterial baroreflex-mediated changes in peripheral vascular tone and heart rate (HR) (Fadel et al. 2003b). As the drop in arterial pressure associated with thigh cuff release does not alter central venous pressure (and hence stroke volume) (Fadel et al. 2001) reflex tachycardia transiently augments cardiac output to the extent that is proportional to the degree of tachycardia. That increases in cardiac output can elevate CBF independent of blood pressure (Ogoh et al. 2005a, 2007) raises the possibility that baroreflex-mediated changes in cardiac output can alter cerebrovascular tone and potentiate dynamic CA. This relationship would have myriad implications with respect to data interpretation in a range of previous studies because typically the baroreflex and CA are held as separate entities, and data has historically been treated under this paradigm. For example, although it has been demonstrated that cardiovascular disease impairs dynamic CA (Paulson et al. 1990; Immink et al. 2004), the impairment may be due to baroreflex dysfunction (Head, 1994; Mussalo et al. 2002) rather than specific dysfunction of cerebrovascular regulation.

Therefore, this study tested the hypothesis that acute baroreflex dysfunction attenuates dynamic CA as assessed using the RoR index. We used combined β-1 adrenergic (metoprolol tartrate) and muscarinic cholinergic (glycopyrrolate) blockade to reduce the baroreflex response.

Methods

Subjects

Nine healthy individuals (7 men and 2 women; aged 26 ± 3 years (mean ±s.d.)) volunteered for this study, which was approved by the New Zealand Health and Disability Ethics Committee (Lower South Regional) and conformed to the standards set by the Declaration of Helsinki. All subjects were free of any known cardiovascular or respiratory diseases and were not taking any medications. Each subject received a verbal and written explanation of the study objectives, measurement techniques, and the risks and benefits associated with the investigation. Prior to the actual experimental day, each subject was familiarised with the equipment and the study protocol. The subjects were requested to abstain from caffeinated beverages for 12 h and strenuous physical activity and alcohol intake for at least 24 h prior to testing.

Measurements

MCA V was measured with a 2 MHz pulsed Doppler ultrasound system (DLW Doppler, Sterling, VA, USA). Beat-to-beat arterial blood pressure (ABP) was monitored with finger photoplethysmography (Finometer, TNO-TPD Biomedical Instrumentation). In five subjects, ABP was also measured by a catheter (20-guage, BD Insyte) placed into the radial artery of the non-dominant arm and connected to a transducer (Deltran II, Utah Medical Products Ltd, RoI). HR was monitored using 3-lead electrical cardiogram. Stroke volume and cardiac output were calculated from the ABP waveform with the Modelflow method, incorporating age, sex, height and weight (BeatScope 1.0 software; TNO-TPD Biomedical Instrumentation). Subjects breathed through a leak-free respiratory mask (Hans-Rudolph 8980, Kansas City, MO, USA) attached to a one-way non-rebreathing valve (Hans-Rudolph 2700). Expiratory flow was measured with a heated pneumotach (Hans-Rudolph HR800). End-tidal () and () were sampled from the mask and measured by a gas analyser (model CD-3A, AEI Technologies, Pittsburgh, PA, USA). Ventilatory and gas values (flow, tidal volume, frequency) were displayed in real time. Expiratory volume was calculated with the integrated flow signal and the frequency of breathing. All data were acquired continuously at 200 Hz using an analog-to-digital converter (Powerlab/16SP ML795; ADInstruments) interfaced with a computer and were analysed using commercially available software (Chart version 5.5.5, ADInstruments).

Experimental protocols

Control

Data were recorded for 5 min prior to 6 min of lower limb blood flow occlusion using thigh cuff inflation (220 mmHg). Rapid deflation (<1 s) of the thigh cuffs occurred at 6 min with measurements recorded for an additional 3 min. Subjects were instructed to breathe according to the sound of a metronome and, if needed, adjust their baseline tidal volume level in order to maintain at baseline levels.

Intervention

After baseline data collection, subjects rested for 30 min before pharmacological intervention. β-1 adrenergic blockade was achieve with step-wise intravenous injections of metoprolol tartrate at a rate of 1 mg min⁻¹ up to the maximum calculated dose (0.2 mg kg⁻¹). Twenty minutes following β-1 adrenergic blockade, the thigh cuff protocol was repeated. Subjects then rested for 30 min before cardiac vagal blockade with step-wise injections of 0.2 mg of glycopyrrolate, continued until mean HR did not change with consecutive 0.2 mg doses. Full cardiac autonomic blockade (metoprolol and glycopyrrolate) was considered complete when sequential bolus injections of sodium nitroprusside (50 μg) followed 60 s later by phenylephrine hydrochloride (50 μg) produced < 5 beats min⁻¹ changes in HR below or above baseline HR. At baseline, bolus injections of sodium nitroprusside and phenylephrine hydrochloride elicited changes in HR greater than ± 25 beats min⁻¹. The thigh cuff protocol was then repeated under conditions of full cardiac vagal blockade. Additional 0.1 mg doses of glycopyrrolate were given as required if mean HR decreased below 5 beats min⁻¹; the extent of the cardiac autonomic blockade was further confirmed following the thigh cuff protocol.

Data analysis

Beat-to-beat mean arterial pressure (MAP) and mean MCA V (MCA V_mean) were obtained from each waveform. A cerebrovascular conductance index (CVCi) was calculated by dividing MCA V_mean by MAP and used as an estimate of changes in cerebrovascular conductance. The derived CVCi during acute hypotension is not directly related to steady-state cerebrovascular conductance because changes in the vascular compliance and resistance affect CBF during dynamic regulation.

Cardiovascular responses to thigh cuff release

The responses of HR, MAP and MCA V_mean following cuff release were identified. Control values of HR, MAP and MCA V_mean were defined by calculating their averages during the 4 s immediately before thigh cuff release. The tachycardic response was calculated by subtracting the control HR value from peak HR, while the MAP and MCA V_mean responses were calculated by subtracting the control values from nadir value at time 1.0 to 3.5 s from cuff release.

Dynamic CA

Control values for CVCi were defined by calculating the 4 s average immediately before thigh cuff release. Changes in MAP, MCA V_mean and CVCi during cuff release were determined relative to their concomitant control values. At time 1.0 to 3.5 s from cuff release the rate of change in CVCi is directly related to CA (Aaslid et al. 1989). The rate of regulation (RoR) was calculated as an index of dynamic CA from the slope of the regression line between CVCi and time (T), normalised by the cuff release-induced hypotension (Aaslid et al. 1989):

where ΔCVCi/ΔT is the slope of the linear regression between CVCi and T; and ΔMAP, the magnitude of the step, was calculated by subtracting control MAP from averaged MAP during the interval from 1.0 to 3.5 s (Aaslid et al. 1989).

Statistics

Statistical comparison of physiological variables and RoR were made utilizing a repeated-measures one-way analysis of variance (ANOVA). A Student–Newman–Keuls test was employed post hoc when interactions were significant. The relationship between two factors was described using linear regression analysis. Statistical significance was set at P < 0.05, and results are presented as means ±s.e.m. Analyses were conducted using SigmaStat (Jandel Scientific Software, SPSS Inc., Chicago, IL, USA).

Results

Intra-arterial measurements were collected in five out of the nine subjects. In the remaining four subjects, ABP was recorded using the Finometer. Manual blood pressure recordings were intermittently used to confirm the accuracy of the finger photoplethysmography measurements. Compared to control, resting HR decreased 9 ± 1 beats min⁻¹ (P= 0.052) during β-1 adrenergic blockade (metoprolol) and increased 28 ± 5 beats min⁻¹ (P < 0.001) during full cardiac autonomic blockade (glycopyrrolate plus metoprolol) (Table 1). In addition, full cardiac autonomic blockade decreased stroke volume (11 ml, P= 0.027), while it increased cardiac output (1.8 l min⁻¹, P= 0.027). However, MAP (P= 0.923), MCA V_mean (P= 0.675), (P= 0.741) and CVCi (P= 0.777) were unchanged under both conditions compared to control.

Table 1.

The haemodynamic responses to β-1 adrenergic and full cardiac blockade

	MAP (mmHg)	MCA Vmean (cm s−1)	HR (beats min−1)	SV (ml)	Q (l min−1)	(mmHg)	CVCi (cm s−1 mmHg−1)
CON	85 ± 4	73 ± 4	67 ± 4	103 ± 6	6.8 ± 0.6	39.7 ± 1.0	0.866 ± 0.043
MET	85 ± 5	70 ± 4	58 ± 4	105 ± 6	6.0 ± 0.5	39.8 ± 0.8	0.831 ± 0.031
GLY	86 ± 4	72 ± 4	95 ± 2*†	92 ± 8*†	8.6 ± 0.7*†	39.1 ± 1.2	0.850 ± 0.053

Value are means ±s.e.m. MAP, mean arterial pressure; MCA V_mean, middle cerebral artery mean blood velocity; HR, heart rate; SV, stroke volume; Q, cardiac output; , partial pressure of end-tidal carbon dioxide; CVCi, cerebral vascular conductance index; CON, control condition; MET, metoprolol condition; GLY, glycopyrrolate with β-1 adrenergic blockade (full cardiac autonomic blockade condition).

^*Different to CON, P < 0.05.

^†Different to MET, P < 0.05.

Thigh cuff release

Thigh cuff release elicited an acute decrease in ABP during all conditions (Fig. 1). Changes in MAP from baseline to nadir were −21 ± 2%, −24 ± 2% and −27 ± 2% during control, β-1 adrenergic and full cardiac autonomic blockade conditions, respectively (Fig. 2). These decreases in ABP were sufficient to evoke a transient decrease in MCA V_mean (−17 ± 2%, −16 ± 2% and −21 ± 2% during control, β-1 adrenergic and full cardiac autonomic blockade conditions, respectively). In addition, the thigh cuff release and subsequent hypotension elicited the anticipated tachycardia. The tachycardia response to acute hypotension under full cardiac autonomic blockade was abolished (+1.2 ± 0.5 beats min⁻¹) compared to both control (+9.4 ± 3.0 beats min⁻¹, P= 0.029) and β-1 adrenergic blockade (+8.0 ± 2.9 beats min⁻¹, P= 0.031). Moreover, in the full cardiac autonomic blockade condition, thigh cuff release induced greater reductions in both absolute MAP (P= 0.004) and MCA V_mean (P= 0.025) compared to control or β-1 adrenergic blockade conditions.

Figure 1. Recordings of arterial blood pressure (ABP), middle cerebral artery blood velocity (MCA V) and heart rate (HR) during thigh cuff release without (A) and with (B) full cardiac autonomic blockade (metoprolol with glycopyrrolate) in one subject.

The thigh cuffs were released at time 0. Dotted line shows mean value.

Figure 2.

A, the rate of regulation (RoR; upper) and heart rate (HR; lower panel) response to thigh cuff release. B, mean arterial pressure (MAP; upper) and middle cerebral artery mean blood velocity (MCA V_mean; lower panel) responses to thigh cuff release during control, β-1 adrenergic and β-1 adrenergic with vagal blockade conditions. All data were taken as averages during 1 to 3.5 s following thigh cuff release. CON, control condition; MET, Metoprolol condition; GLY, glycopyrrolate with β-1 adrenergic blockade (full cardiac autonomic blockade condition). *Different to CON, P < 0.05. #Different to MET, P < 0.05.

The RoR was attenuated in the full cardiac autonomic blockade condition (0.255 ± 0.021 s⁻¹) compared with the control and β-1 adrenergic blockade conditions (0.353 ± 0.033 s⁻¹, P= 0.028 and 0.384 ± 0.047 s⁻¹, P= 0.015). In addition, the RoR was related to the tachycardia response (n= 27 (pooled data), r= 0.7; P= 0.015).

Discussion

The main findings of this study are: (1) dynamic CA, as indexed by the RoR, was attenuated following combined β-1 adrenergic and muscarinic cholinergic blockade, and (2) the RoR attenuation was related to the reduction in tachycardia following autonomic blockade. These findings support our hypothesis that acute cardiac baroreflex dysfunction attenuates dynamic CBF regulation.

Aaslid et al. (1989) used TCD and rapid thigh cuff deflation to non-invasively and non-pharmacologically evaluate dynamic CA in resting humans. This technique, however, cannot be used under several conditions e.g. during dynamic leg cycling exercise. Dynamic CA is a frequency-dependent phenomenon (Zhang et al. 1998); therefore, transfer function analysis has been used (Zhang et al. 2002; Ogoh et al. 2005b,c;) to quantify dynamic CA during various conditions including exercise. Studies employing either technique have only assessed the MCA V responses to a rapid and transient drop in ABP, potentially neglecting other regulating factors during dynamic CBF regulation. A transient change in ABP can load or unload the arterial baroreceptors (Fadel et al. 2003a) triggering a reflex change in HR via the arterial baroreflex (Fadel et al. 2003b).

Peak cardiac baroreflex responses occur within 2–3 s following stimulation onset (Potts & Raven, 1995), which overlaps the time used for calculating the RoR index (1 to 3.5 s) (Aaslid et al. 1989). This overlap is potentially important because baroreflex-mediated HR changes can alter cardiac output, which has been shown to be an independent determinant of MCA V_mean in humans (Ogoh et al. 2005a, 2007). Furthermore, given that thigh cuff deflation does not alter central venous pressure (Fadel et al. 2001), transient increases in cardiac output, and resultant increases in CBF, probably parallel the degree of tachycardia. Indeed, in the present study, full cardiac autonomic blockade was associated with attenuated RoR (Fig. 3), which was related to the reduction in tachycardic responses (P= 0.015). These findings support our hypothesis that acute blunting of the cardiac baroreflex attenuates dynamic CBF regulation, and indicates that the ‘CA response’ identified by the thigh cuff release technique reflects the integrated activity of both the cardiac baroreflex and dynamic CA.

Figure 3.

Normalised beat-to-beat continuous recordings of heart rate (A; HR) and cerebral vascular conductance index (B; CVCi) during thigh cuff release with (β-1 adrenergic blockade with glycopyrrolate (GLY); dotted line) and without full cardiac autonomic blockade (control (CON); continuous line) in one subject. The thigh cuffs were released at time 0. Straight lines through CVCi data were determined by linear regression analysis from 1 to 3.5 s (black bar) after cuff release. All data are shown in normalised units relative to control pre-release values obtained during −4 to 0 s.

It is well established that the arterial baroreflex has two components – cardiac and vasomotor reflexes – that buffer moment-to-moment fluctuations in arterial blood pressure (Fadel et al. 2003a). Ogoh et al. (2003) reported the different contributions of cardiac and vasomotor reflexes on arterial blood pressure regulation; they concluded that alterations in vasomotor activity are the primary means by which the arterial baroreflex regulates ABP, because the contribution of cardiac output response to peak blood pressure was only 23%. However, at the acute phase (2–3 s from the start of stimulation) the contribution of cardiac output to blood pressure regulation was almost 100%, despite a small change in blood pressure. Similarly, in the present study during full cardiac autonomic blockade, an attenuated tachycardia response exacerbated the cuff release-induced reductions in MAP (P= 0.004), and consequently evoked a greater transient decrease in MCA V_mean (Fig. 2). These findings indicate that the baroreflex-induced tachycardia response following acute hypotension regulates the reduction in ABP, thus acting to minimise decreases in CBF (as indexed by MCA V_mean).

Perspectives

Epidemiological studies have clearly established that impairment of the arterial baroreflex is associated with increased mortality risk from stroke (La Rovere et al. 1998, 2002). However, despite clear prognostic associations, it remains unclear whether impaired baroreflex function is the cause or rather the consequence of stroke. Using a hypertensive rat model of stroke, Liu et al. (2007) recently showed that impaired baroreflex function was an independent predictor of stroke incidence. Moreover, it was shown that restoration of baroreflex function with the antihypertensive Ketanserin was protective against stroke even without blood pressure correction, indicating that the integrity of baroreflex function per se predicts the risk of stroke. The results of the present study indicate that acute baroreflex impairment attenuates dynamic CBF regulation, thus illustrating one potential mechanism by which diminished baroreflex function may increase stroke risk. However, we also acknowledge that whilst acute baroreflex dysfunction attenuated dynamic CA, our findings may not extend to situations where baroreflex function is declining gradually or persists chronically. For example, dynamic CA is preserved despite progressive baroreflex impairment with age (Carey et al. 2000) while spontaneously hypertensive patients with baroreflex dysfunction have better dynamic CA compared to normotensive controls (Serrador et al. 2005). These findings indicate that the relationship between baroreflex function and dynamic CA is not necessarily additive but that underlying neuroplasticity may enable dynamic CA to undergo compensatory adjustments in the presence of chronic or slowly progressing baroreflex dysfunction. Such interactions, which were not explored in the present study, clearly warrant further investigation.

Limitations

Potential limitations of the present study should be considered. TCD measures CBF velocity in the MCA rather than CBF; however, the MCA diameter changes little with acute haemodynamic perturbations (Giller et al. 1993) such as those elicited during the thigh cuff release protocol (Aaslid et al. 1989). Therefore, we are confident that TCD measures of the transient changes in MCA V_mean reflect those of the transient changes in CBF (Aaslid et al. 1989; Giller et al. 1993; Serrador et al. 2000; Secher et al. 2008). In addition, it is difficult to identify the contribution of cerebral vasodilatation (vasomotion) or systemic control to the dynamic CBF regulation. Another consideration is that the tachycardia response to acute hypotension was calculated from a data collection period that only lasts 1 to 3.5 s. However, the time to peak change in HR baroreflex response occurs within 2–3 s from the start of stimulation (Potts & Raven, 1995) and therefore reflects the key period in which baroreflex-mediated changes in HR may affect dynamic CA.

Conclusion

In summary, acute blunting of the cardiac baroreflex attenuates dynamic CBF regulation, indicating that the cardiac baroreflex plays an important role in the dynamic CBF regulation during acute transient hypotension.

Glossary

Abbreviations

ABP

arterial blood pressure

CA

cerebral autoregulation

CBF

cerebral blood flow

CVCi

cerebrovascular conductance index

HR

heart rate

MCA

middle cerebral artery

RoR

rate of regulation

TCD

transcranial Doppler

V_mean

mean blood velocity