|Home | About | Journals | Submit | Contact Us | Français|
To compare changes in cerebral autoregulation in response to controlled, lower body negative pressure‐induced hypotension in patients with carotid sinus syndrome (CSS) and case controls.
Prospective case controlled study.
Secondary and tertiary referral falls and syncope service.
17 consecutive patients with CSS and 11 asymptomatic controls.
Hypotension insufficient to cause syncope induced by lower body negative pressure (minimum 30 mm Hg fall in systolic blood pressure (SBP)) during concomitant transcranial Doppler ultrasonography.
Cerebral autoregulation (systolic, diastolic and mean middle cerebral arterial blood flow velocities and cerebrovascular resistance) with continuous end‐tidal carbon dioxide and haemodynamic monitoring.
Cerebral autoregulatory indices differed significantly between patients with CSS and controls. Systolic, diastolic and middle cerebral arterial blood flow velocities were, respectively, 9.2 m/s (95% confidence interval (CI) 2.9 to 15.4 m/s), 4.7 m/s (95% CI 1.5 to 7.9 m/s) and 6.9 m/s (95% CI 2.5 to 11.4 m/s) slower in patients with CSS. Cerebrovascular resistance was significantly greater in patients with CSS than in controls at SBP nadir and suction release; differences were 0.9 mm Hg/m/s (95% CI 0.0 to 1.7 mm Hg/m/s) and 0.8 mm Hg/m/s (95% CI 0.0 to 1.7 mm Hg/m/s), respectively. End‐tidal carbon dioxide and systemic haemodynamic variables were similar for patients and controls at baseline and during lower body negative pressure.
Cerebral autoregulation is altered in patients with CSS. This difference may have aetiological implications in the differential presentation with falls and drop attacks rather than syncope.
Carotid sinus syndrome (CSS) is among the most common causes of syncope among older people, being causal in up to one quarter of those older than 70 years.1 Recent evidence has shown that patients with CSS may present with syncope or falls,2,3,4 with cardiac pacing successfully reducing the frequency of both falls and syncope in the syndrome.2,3 Amnesia for loss of consciousness is a well recognised characteristic of the syndrome and has been reproduced in experimental studies in up to 80% of patients with witnessed loss of consciousness during carotid sinus massage‐induced asystole.4 Falls are most often attributed to amnesia for loss of consciousness during syncope in the context of unwitnessed events. One possible explanation for this amnesia for loss of consciousness is an underlying derangement of cerebral autoregulation. Although cerebral autoregulation has been measured with a variety of techniques, transcranial Doppler ultrasonography (TCD) has become the most often used approach in recent years because of its non‐invasive nature, relative simplicity and accuracy.5,6,7,8,9,10,11,12,13,14,15,16,17 TCD has been used to examine cerebral autoregulation in response to hypotension both in healthy subjects9,11,12,13,14 and in patients with vasovagal syncope,10,11,12,13,14,15 but there are few data on cerebral autoregulation in patients with CSS.16,17 To study this phenomenon without the catastrophic cerebral autoregulation derangement consequent on profound asystole and hypotension induced by carotid sinus massage in patients with CSS, a more graded fall in systemic blood pressure can be induced by lower body negative pressure (LBNP), a technique used originally to study the effects of profound orthostasis on healthy people.18 The technique is now used in the diagnosis of vasovagal syncope19,20 and avoids the unnecessary risks associated with carotid sinus massage.21 Indeed, LBNP testing has previously been used to provoke experimental hypotension to evaluate cerebral haemodynamic function in healthy subjects12,13,14 and those with vasovagal syncope,15 but never in patients with CSS.
We hypothesised that patients with CSS have different cerebral autoregulatory control from case controls. The objective of this study was thus to compare changes in cerebral autoregulation in response to controlled, LBNP‐induced hypotension in patients with CSS and asymptomatic controls.
Consecutive patients in sinus rhythm with CSS as the sole attributable cause of symptoms (as determined by investigation according to internationally recognised guidelines22,23) presenting to our tertiary referral falls and syncope service were invited to participate in the study. CSS was of the cardioinhibitory subtype and was diagnosed in patients with recurrent events who had asystole in excess of 3 s during carotid sinus massage.19 Control subjects were of similar age, sex and co‐morbidity, had no history of falls, dizziness or syncope and were recruited through a database of control subjects. Drugs were discontinued a minimum of five half lives before testing. All participants gave fully informed, written consent. The study had approval from the Newcastle and North Tyneside local research ethics committee.
Cerebral blood flow velocity was assessed with TCD following the methods of Lipsitz et al.13 After a temporal ultrasonic bone window was confirmed, the right middle cerebral artery was insonated with a 2 MHz TCD probe (Scimed, Bristol, UK), which was fixed in place by a Mueller–Moll probe fixation device. Spectral signals were observed continuously to ensure good signal quality. Systolic (SBFV), diastolic (DBVF) and mean cerebral blood flow velocities (MBFV) were recorded continuously, with cerebrovascular resistance (CVR) derived from the formula CVR =MAP/MBFV, where MAP is mean arterial blood pressure.13 Resistance and pulsatility indices (derived measures relating SBFV, DBFV and MBFV: resistance index = SBFV − DBFV/SBFV; pulsatility index = SBFV − DBFV/MBFV) were also calculated. Because the patient was supine during the procedure, correction for hydrostatic pressure differences between heart and brain was unnecessary.
The lower half of the participant (to the level of the iliac crests) was enclosed in an airtight chamber with a suction engine attached, which allowed calibrated negative pressure to be applied up to 90 mm Hg. The predetermined required systolic blood pressure (SBP) nadir of 30 mm Hg was achieved by supine, graded LBNP. Beat‐to‐beat SBP, diastolic blood pressure (DBP), MAP and heart rate (surface ECG at 25 m/ms) were recorded continuously through digital photoplethysmography (Finapres, Ohmeda, Wisconsin, USA).24
Subjects lay supine in a dimly lit, quiet room at a constant temperature for 15 min before the procedure. ECG and blood pressure were monitored throughout, as was end‐tidal carbon dioxide concentration (Etco2; through a close‐fitting face mask and infrared capnography), to monitor potential confounding from the effects of swings in arterial carbon dioxide concentrations on cerebral arteriolar diameter.25,26 After insonation of the middle cerebral artery and fixation of the ultrasound probe, all baseline parameters were recorded at rest (baseline) and at the four time points described in table 11 and shown graphically in fig 11.. The relatively steady‐state blood pressure changes during baseline, noise, nadir (because of the graded fall in blood pressure over several minutes) and return to baseline provide a measure of static cerebral autoregulation as described by Paulson et al,5 whereas the rapid change in blood pressure over a few seconds at overshoot is more indicative of dynamic cerebral autoregulation.5
Data were recorded continuously with LabWindows software. To evaluate the beat‐to‐beat variations in systemic blood pressure, heart rate, Etco2 and cerebral blood flow velocities, an observer blinded to patient status examined the raw wave forms for each patient at each time point. The values derived at each time frame were averaged from the five beats around each point and the resulting values were averaged for each group as a whole.9,13 An independent assessor repeated this procedure in a random sample of five patients and controls to ensure agreement.
The comparison between CSS and controls was the primary objective of the research. Continuous variables (age, blood pressure) were plotted to check for outliers and wildly skewed distributions. The distribution of data was normal with the exception of RR interval responses to carotid sinus massage. Statistics are reported as mean (SD) for all comparisons with the exception of RR interval response to carotid sinus massage, which is reported as median (extreme range). Because the data are continuous variables, groups were compared at a single time point by independent sample t tests and over the five time points—baseline, noise, nadir, overshoot and return to baseline—by repeated measures analysis of variance. Data were analysed with the SPSS V.10.2 software package (SPSS Inc, Chicago, Illinois, USA).
Eighteen patients with CSS completed the study but data were uninterpretable for one because of interference from an essential tremor. Of 12 controls, a suitable temporal bone ultrasound window was not obtainable for one. Age, sex or co‐morbidity did not differ between patients and controls (table 22).). Patients with CSS had a mean 5.6 (SD 1.8) s asystole on carotid sinus massage.
LBNP at SBP nadir was similar in both control subjects and patients (35.2 (SD 7.8) v 35.7 (SD 10.1) mm Hg, p = 0.89). No participant experienced presyncope or syncope. Figure 22 shows the distribution of responses for patients and controls for nine haemodynamic and cerebral autoregulatory indices. SBP, DBP, heart rate and Etco2 were similar for patients and controls at each time point: baseline, SBP nadir, SBP overshoot and return to baseline. MAP and heart rate followed a similar pattern, with Etco2 remaining relatively constant (fig 22,, table 33).
Table 44 shows the results of repeated measures analysis of variance. For each index there is an overall test of differences between time points (baseline, noise, SBP nadir, overshoot and return to baseline), an overall test of differences between groups (patients with CSS and controls) and a test of an interaction effect (the differences between time points were different for patients with CSS and controls).
For the blood pressure variables, the differences between time points were highly significant but differences between patients with CSS and controls were not. For DBP, although overall the groups did not differ, there was some evidence of an interaction between group and time point; however, post hoc t tests indicated that the difference in means was not significant at any time point.
For the blood flow velocity variables, differences between time points were not as notable (and were significant only for SBFV) but patients with CSS and controls differed significantly: the blood flow velocity was considerably less for patients with CSS ((tablestables 3 and 44).). The differences between groups were 9.2 m/s (95% confidence interval (CI) 2.9 to 15.4 m/s), 4.7 m/s (95% CI 1.5 to 7.9 m/s) and 6.9 m/s (95% CI 2.5 to 11.4 m/s) in SBFV, DBFV and MBFV, respectively. For CVR there were significant differences between time points, significant differences between patients with CSS and controls, and a significant interaction term ((tablestables 3 and 44,, fig 11).). The difference between patients with CSS and controls was small at baseline (0.3, 95% CI −0.3 to 1.0) and larger at noise (0.7, 95% CI −0.1 to 1.5), SBP nadir (0.9, 95% CI 0.0 to 1.7), overshoot (0.8, 95% CI 0.0 to 1.6) and return to baseline (0.8, 95% CI 0.0 to 1.7). Pulsatility and resistance indices were similar in both groups at all five time points (table 33).).
The significant differences in MBFV throughout the study and those in CVR on the SBP nadir, overshoot and return to baseline time points provide compelling evidence for disordered cerebral autoregulatory mechanisms in CSS. In the related neurally mediated disorder vasovagal syncope, analyses of several case series have shown a rise in CVR during head‐up tilt‐induced presyncope and syncope,12,24,27 but later work related this apparently paradoxical cerebral vasoconstriction to hyperventilation‐induced hypocapnia.28 Our findings echo those of Grubb et al,27 Levine et al12 and Bondar et al,14 showing significantly higher CVR in patients with CSS at SBP nadir, overshoot and return to baseline despite a highly significant average MBFV 23% lower in patients than in controls. In the presence of normocapnia (as in our study population) the converse is expected. In previous studies, cerebral blood flow velocity and carbon dioxide concentration fell during LBNP sufficiently to induce presyncope and syncope (with a potentially causal relationship between the two), but we observed neither of these changes in our study, presumably reflecting the relative modesty of the LBNP stimulus essential to our study design.
It has been argued that augmented sympathetic nervous activity during LBNP causes cerebral vasoconstriction, which overwhelms the usual mechanisms promoting vasodilatation in the face of a fall in MBFV.12,14,27 Indirect support for this contention came from a recent study of healthy young subjects examined by TCD during LBNP and ganglion blockade, in which Zhang et al29 elegantly showed that the cerebral circulation is likely to be under tonically active autonomic neural control. Although cerebral autoregulation was maintained in both patients with CSS and controls in response to LBNP, the wide differences between the groups in the variables described above suggests a baseline difference in autoregulation in those with CSS. Data are emerging showing a high prevalence of cognitive impairment and dementia in patients with CSS.30 This is attributed to small vessel disease reflected in a higher prevalence of white matter lesions in patients with CSS, with the density of white matter lesions on magnetic resonance imaging correlating with the degree of carotid sinus massage‐induced hypotension.31 We hypothesise that the abnormalities in cerebral autoregulation at baseline are due to such microvascular disease, manifesting as abnormal cerebral autoregulation. Our results raise the intriguing possibility that patients with CSS are prone to relative (and paradoxical) tonic intracerebral vasoconstriction, which predisposes them to further inappropriate vasoconstriction during CSS‐mediated vasodepression and asystole, when vasodilatation should otherwise supervene. We speculate that such paradoxical vasoconstriction preferentially affects areas intimately related to short term memory and consciousness level, with the clinical corollary being CSS presenting as falls.
Although TCD during carotid sinus massage would arguably provide a more realistic assessment, we deliberately avoided this approach because of the possibility of neurological complications during repeated carotid sinus massage21 and the confounding influence of unilateral cerebral artery occlusion on CVR in the contralateral side. Others have done so, however, and in two small, uncontrolled series Leftheriotis et al16,17 examined cerebral blood flow velocity in patients with CSS during carotid sinus massage after 10 min in the head‐up tilt position. Cerebral autoregulation failed only during a fall in systemic blood pressure below the lower limit for normal cerebral autoregulatory function—that is, 50 mm Hg32—with cerebral perfusion falling by 50% and CVR rising (non‐significantly) during carotid sinus massage, then falling rapidly immediately afterwards.16,17 These results are consistent with those presented above and with the vasovagal data of Carey et al.10
As with all TCD assessments of cerebral autoregulation, the technique is dependant on several assumptions. Firstly, middle cerebral artery cross‐sectional diameter is assumed to be constant. Secondly, as the Doppler principle depends on velocity, derivative measurement of blood flow relies on a linear relationship between flow and velocity. These factors have proved controversial, with some early criticism of the technique on the basis of the assumptions made.7,8,32 Later research comparing direct measures of internal carotid artery flow (during carotid surgical procedures) with middle cerebral artery velocity measured by TCD found that TCD flow measurements accurately mirrored changes in internal carotid arterial flow.9 Others similarly showed that changes in middle cerebral artery velocity correlated with cerebral blood flow,8,33,34 although with the caution that absolute velocity should not be used as an indicator of cerebral blood flow.8 Nonetheless, as middle cerebral arterial diameter was not measured as part of the study, this cannot be absolutely excluded. Furthermore, although the derivation of CVR from indirect measures of systemic blood pressure can be criticised, Zhang et al29 recently found a close correlation between digital photoplethysmographically measured blood pressure and intra‐arterial pressure during LBNP‐related orthostatic stress.
Cerebral autoregulation is abnormal in patients with CSS and may explain some of the clinical features of the disorder. Additional studies, with power calculations based on this work and by using TCD during more profound hypotension (induced by tilt, LBNP or carotid sinus massage‐induced asystole), may help explore this issue further.
The authors acknowledge the support of the British Heart Foundation for this work.
CSS - carotid sinus syndrome
CVR - cerebrovascular resistance
DBFV - diastolic blood flow velocity
Etco2 - end‐tidal carbon dioxide concentration
LBNP - lower body negative pressure
MAP - mean arterial blood pressure
MBFV - mean cerebral blood flow velocity
SBFV - systolic blood flow velocity
SBP - systolic blood pressure
TCD - transcranial Doppler ultrasonography
Sources of funding: Dr Parry was supported by the British Heart Foundation.
Conflicts of interest: None declared.