Cerebrovascular Function in Aging and Dementia: A Systematic Review of Transcranial Doppler Studies Open Access

Aging and dementia are associated with changes in cerebrovascular structure and function which contribute to associated cognitive and functional declines [1,2,3]. Recent autopsy studies have stressed the important role of vascular pathologies in the manifestation of late-onset dementia [4,5,6]. Further, systematic reviews and meta-analyses have highlighted the importance of vascular risk factors (e.g. hypertension and stroke) on dementia onset and progression [7,8,9,10]. Research investigating functional cerebrovascular contributions to cognitive performance in age-related decline and dementias such as Alzheimer’s disease (AD) and vascular dementia (VaD) have generally reported reduced cerebral perfusion [11,12]. These studies have employed techniques such as functional magnetic resonance imaging, positron emission tomography or single-photon emission computed tomography. However, these techniques are expensive, and there are feasibility issues which are particularly problematic for older populations, including the need for individuals to be sufficiently mobile to attend a research facility, lie still for a prolonged duration and have no metal implants.

Transcranial Doppler (TCD) ultrasonography is a non-invasive, inexpensive and portable technique with high temporal resolution, allowing continuous and bilateral recording of cerebral blood flow velocity through the major arteries (e.g. medial, anterior, posterior and basilar). Measurements can be taken at rest, during hypercapnia or hypocapnia (to assess cerebrovascular reactivity), or during cognitive tasks. TCD data collected during cognitive operations is commonly referred to as functional TCD (fTCD) and is the assessment of blood flow velocity change in response to a specific cognitive stimulus or mental operation. Resultant graphs displaying blood flow velocity versus time are known as evoked-flows and are generated in a similar manner to event-related potentials (ERPs) derived from electroencephalogram data where multiple trials are averaged relative to the presentation of cognitive stimuli [13]. Pioneered in the 1980s [14], there has been a recent resurgence in the use of the TCD technology, particularly in the aging and dementia fields.

Haemodynamic abnormalities may be critical markers of dementia and cognitive decline in elderly individuals. Chronic cerebral hypoperfusion could affect cellular health within the brain and the development of neurodegenerative pathologies [15,16]. TCD methods have much to provide the assessment of functional cerebrovascular contributions to cognitive impairment in dementia and aging and may help in the differentiation of dementia from normal aging and between the subtypes such as AD and VaD. This paper aims to systematically review previous research assessing dementia [including mild cognitive impairment (MCI), an intermediate state between normal aging and dementia] and aging using TCD techniques, and, in doing so, summarise key protocols, metrics and consistent findings to point to areas of future research.

The PubMed and PsychInfo databases were searched on March 19, 2012, using the search terms: (‘transcranial doppler’) AND (dementia OR age OR ageing OR aging OR Alzheim* OR ‘mild cognitive impairment’). A total of 1,172 articles were retrieved. Titles and abstracts were read by at least two of the authors. Articles were retained if they collected TCD data at rest, hyper-/hypocapnia or during a cognitive task from any cerebral artery accessible via TCD [the anterior cerebral artery (ACA), middle cerebral artery (MCA), posterior cerebral artery (PCA) or basilar artery] and assessed dementia (differences between subtypes or with normal aging; or changes associated with clinical progression) or aging in those over 50 years of age. Articles were included if they assessed general late-onset dementia (i.e. without subtype classification) or AD and VaD (and its historical counterpart multi-infarct dementia) subtypes. Articles were excluded if they were not written in English or were case reports. Articles with TCD as an outcome in a clinical trial of a pharmaceutical compound were also excluded unless they presented baseline (i.e. prior to drug administration schedule) comparison data (e.g. with a healthy control or another subtype of dementia).

Details from each included study are summarised in table 1 and include: study sample (including sex, age and diagnoses), TCD protocol, vessel(s) investigated, TCD metrics analysed and key findings.

The calculation of common resting TCD metrics employed are summarised in figure 1, including systolic peak flow velocity, end diastolic velocity, mean flow velocity (MV) and the pulsatility index and resistance index. Cerebrovascular reactivity to hypocapnia (reduced CO₂ in blood) or hypercapnia (increased CO₂ in blood) was calculated as the difference between mean flow velocity during hypo-/hypercapnia and resting mean flow velocity, divided by resting mean flow velocity [i.e. (MV during capnia – resting MV)/resting MV]. The periods of time where resting or hypo-/hypercapnia mean flow velocities were measured varied between studies. The breath-holding index (BHI) was also commonly used to assess velocity changes in response to hypercapnia which was calculated as the difference between mean flow velocity at the end of a breath hold (usually at least 30 s) and resting mean flow velocity, divided by resting mean flow velocity [i.e. (MV at end of breath hold – resting MV)/resting MV].

Thirty-four articles were selected for review: 29 assessed TCD measures during rest, 13 during hyper-/hypocapnia and 4 employed fTCD (cognitive) techniques (some articles presented multiple assessments). The vast majority of articles assessed differences between demented and non-demented groups rather than age-related changes (over 50 years). There was only one population-based sample used, the Rotterdam Study [15,17], which comprised over 1,700 individuals. The remaining studies employed clinic or convenience recruited samples. All studies are summarised in table 1.

A number of TCD metrics were employed to assess cerebral blood flow at rest including mean flow velocity, systolic velocity, diastolic velocity, the pulsatility index and other resistance measures, and the flow asymmetry index (i.e. the difference between right and left arteries at rest). The majority of papers assessed the MCA or PCA. Other vessels investigated included the ACA and the basilar artery. Most papers assessed differences between those with and without dementia or dementia progression. However, some age associations in late life were reported: the pulsatility index increased with age [17,18], while mean, diastolic and systolic flow velocities decreased with age in AD [19] and in the general population [17].

Most studies reported no significant differences in resting mean flow velocities between demented and control groups [20,21,22,23,24,25,26,27,28], or, lower mean flow velocities in demented groups including AD [29,30,31,32,33,34,35], VaD or multi-infarct dementia [30,32,34,36], regardless of subtype [15]. Mean flow velocity measures in AD and VaD groups appeared similar [18,24,30,32,34,36,37], although this was not always the case. Ries et al. [38] reported that the multi-infarct dementia group displayed slower mean and diastolic flow velocities compared to both control and AD groups. Few studies reported flow velocities other than the mean; however, Caamaño et al. [34] reported that individuals with AD and multi-infarct dementia displayed reduced systolic and diastolic (as well as mean) flow velocities compared to healthy controls. The two articles which examined a MCI group assessed mean flow velocity and reported no differences between the MCI group and the control group [29,39], unless restricting to amnestic MCI [29]. One study reported that mean flow velocities were more asymmetric in an AD group as compared to healthy controls [33]. In non-demented controls, lower mean flow velocities were associated with preceding cognitive impairment and smaller hippocampal and amygdala volumes [15].

The pulsatility index was the most commonly employed measure of vessel resistance and was found to be increased in AD [23,29,31,32,36,40] and VaD or multi-infarct dementia patients [25,32,36] as compared to healthy controls, although this was not the case in three papers [24,28,41]. There were many reports that those with VaD or multi-infarct dementia displayed higher pulsatility indexes than those with AD [18,24,28,34], although there were three reports of non-significant differences between subtypes [25,32,36]. There appeared to be no significant difference in the pulsatility index between a MCI group and controls [29]. Employing a different calculation of resistance, van Beek et al. [20] reported increased cerebrovascular resistance in an AD group.

A series of papers by Purandare and colleagues [42,43,44,45,46,47] assessed the significance of spontaneous emboli as detected during resting TCD to dementia and its progression and symptoms. In an article published in 2005, they reported the presence of emboli to be associated with VaD [45]. In a summary of a sequence of studies using the same large cohort [reviewed in [42]], the group went on to report emboli to be associated with both VaD (37% of cases) and AD (40% of cases), as compared to controls (12% of cases). Further, they showed that the presence of emboli was associated with depression symptomatology in dementia and worse cognitive decline over 6 months in controls (i.e. non-demented cases at baseline). Recently, the group has reported emboli to be associated with more rapid cognitive and functional decline in AD and VaD, as well as with an increased number of psychiatric symptoms over a 24-month period [47].

The BHI was commonly employed to assess the effects of hypercapnia [31,48,49,50], along with the administration of gasses with varying %CO₂ [15,32,39] or a pharmaceutical compound [21,37], and closed-circuit rebreathing [27]. In a large population-based sample, cerebrovascular response to hypercapnia was found to significantly decrease with age [17]. Two papers reported that the BHI was lower in AD patients as compared to healthy controls [31,49], similar to Lee et al. [27] who employed closed-circuit rebreathing, while another two papers reported the BHI was only lower in a multi-infarct dementia group (not an AD group) as compared to healthy controls [36,51]. Silvestrini et al. [48] used the BHI to investigate differences between patients with mild/moderate AD with/without carotid stenosis. They reported that those with stenosis were more likely to develop severe dementia over 1 year, and this was related to the BHI ipsilateral to the stenosis. In another paper investigating progression, Silverstrini et al. [50] reported that, in a group diagnosed with AD, lower BHIs were associated with a more rapid cognitive decline.

Vicenzini et al. [32] used a gas mixture (with increased CO₂) to induce hypercapnia, followed by hyperventilation to induce hypocapnia. They reported that the vasomotor range (i.e. taking into account flow velocities during hyper- and hypocapnia) was reduced in AD and VaD patients. Ruitenberg et al. [15] reported no differences in cerebrovascular reactivity using the same protocol between demented (regardless of subtype) and non-demented individuals in a large population-based sample. However, in this case there were only 14 individuals diagnosed with dementia (as compared to 1,730 non-demented subjects). Interestingly, lower cerebrovascular reactivity in the non-demented group was found to be associated with preceding cognitive decline [15]. Anzola et al. [39] reported no differences in cardiovascular response to hypercapnia between MCI and control groups.

Gur et al. [21] and Likitjaroen et al. [37] investigated cerebrovascular reactivity via the administration of acetazolamide, which indirectly induces hypercapnia. Likitjaroen et al. [37] reported no significant differences in cerebrovascular reactivity between AD and VaD groups using this compound (there was no healthy control group). Gur et al. [21] reported no significant differences in cerebrovascular reactivity in individuals who did and did not convert to dementia 3–6 months after a first-ever ischaemic stroke.

No article was identified using fTCD metrics to investigate changes over the age of 50 years. Four fTCD studies compared dementia subtypes and controls: two found that individuals with AD had an attenuated response to cognitive demand [35,52], and two found no differences between AD patients and controls [26,51]. Matteis et al. [51], however, reported that their AD group displayed a reduction in lateralised function, and Asil and Uzuner [26] reported an attenuated response only in their VaD group.

This systematic review revealed 34 articles using TCD methods to investigate aging or dementia. This technique appears to be a feasible method of investigating cerebrovascular function during rest, hyper-/hypocapnia and cognition in old and demented individuals. Measures of vessel resistance and the presence of emboli as detected via TCD appeared the best discriminators of dementia from normal aging. There was also evidence for cerebrovascular reactivity to hypo- and hypercapnia.

Measures of vessel resistance during resting TCD, particularly the pulsatility index, were consistently associated with the presence of dementia, both in AD [20,23,29,31,32,36,40] and VaD or multi-infarct dementia [25,32,36]. Further, the pulsatility index appeared to discriminate between these dementia subtypes [18,24,28,34]. Subjects with VaD appeared to have the highest pulsatility indexes (associated with high vessel resistance), controls the lowest and AD patients sitting in between. Another consistent discriminator between those with and without dementia was the presence of spontaneous cerebral emboli as detected via TCD, although all findings come from one participant group [42,43,44,45,46,47].

Cerebrovascular reactivity to hypo- or hypercapnia was also a good discriminator of dementia, using the BHI [31,49], a gas [32] or closed-circuit rebreathing [27]. Effects may be stronger in VaD [36], similar to pulsatility index findings. Ruitenberg et al. [15] reported cerebrovascular reactivity was a good marker of future cognitive decline in non-demented individuals. However, they reported no differences in cerebrovascular reactivity between demented and non-demented groups using gas stimuli. The sample in this study was large and population based, however lacked the necessary power to detect effects as only 14 individuals were reported to have dementia. In contrast to measures of reactivity to hypo- or hypercapnia, the administration of pharmaceutical compounds to directly induce cerebrovascular reactivity was not as successful in the discrimination of dementia from normal aging and in predicting future decline [21,37], possibly due to their indirect effect and inter-subject metabolic differences.

Findings in relation to differentiating demented and non-demented groups using resting flow velocities were not consistent. Resting flow velocities including mean, systolic and diastolic velocities appeared similar [20,21,22,23,24,25] or lower in demented groups including AD patients [29,30,31,32,33,34,35], VaD/multi-infarct dementia patients [30,32,34,36] or general dementia patients [15], as compared to healthy controls. There appeared to be no significant differences in blood flow velocities between dementia subtypes [18,24,30,32,34,36,37]. Resistance, emboli and cerebrovascular reactivity TCD measures appeared better discriminators of dementia. However, this may be due to the selection of non-optimal resting flow measures. For example, Rosengarten et al. [53] along with Rosengarten and Kaps [54] reported that systolic velocity is less prone to artefacts and more sensitive to the regulation of blood flow than mean flow velocity.

fTCD metrics showed that individuals with dementia had an attenuated response to cognitive load [26,35,52] or a reduction in lateralised function [51] which may reflect compensatory cognitive ability [55,56]. Interestingly, a study of resting TCD reported that mean flow velocities were more asymmetric in an AD group as compared to controls [33] – opposite to that reported in the fTCD paper [51]. Future studies are needed to confirm these findings. There were no identified articles that assessed fTCD changes after 50 years of age.

Abnormalities in TCD measures in demented groups could reflect a number of pathological processes such as cerebral amyloid angiopathy [6,49], arteriolosclerosis or endothelial dysfunction, particularly within the microvascular system [1]. Attenuations in the responsiveness of the cerebrovascular system during cognitive tasks in old and demented subjects may also be a function of surrounding neurons and astrocytes (not signalling for sufficient supply). This abnormal cerebral blood flow may be a cause or consequence of age- and dementia-related neuropathology such as cerebral atrophy. For example, it could simply be that reduced blood flow velocities represent the reduced metabolism of an atrophied brain. Alternatively, reduced blood flow velocities and thus flow may directly lead to cellular dysfunction and death in vulnerable areas such as the hippocampus [15,57]. Ruitenberg et al. [15] found that there was a negative association between resting mean velocity and hippocampal/amygdala volume, and that cerebrovascular disease did not mediate this relationship. This suggests that cerebral blood flow velocity may be directly associated with the volume of brain structures.

Future studies need to investigate neuronal and vascular systems in parallel to assess neurovascular coupling. It is possible to simultaneously record electrical brain activity (via ERPs) and blood flow velocity (fTCD) to assess coupling [52,53,58,59]. The combined ERP-TCD protocol provides an inexpensive, non-invasive method for measuring neurovascular coupling which has the potential to be significantly developed.

One limitation of the TCD method is the assumption that the artery diameter remains constant and therefore any change in velocity represses a change in flow. It has been reported, however, that the diameter of the MCA does not significantly change during moderate alterations in blood pressure (e.g. around 30 mm Hg) [60,61] and therefore any change in velocity reflects a change in blood volume through the artery. Another potential limitation relates to the structure of the temporal window. With age, the temporal window where the TCD ultrasound probe is placed thickens, making recording more difficult. A population-based study reported 25% of participants were lost due to failure to obtain an adequate TCD signal, especially marked in older women [15,17]. These failure rates should be taken into account when planning TCD-based studies. The technique is, however, well tolerated, portable, does not require participants to remain still and allows metal implants to remain in place, unlike expensive and high-spatial resolution cerebral blood flow imaging techniques such as positron emission tomography and single-photon emission computed tomography. Furthermore, TCD equipment is widespread in clinical and research facilities around the world.

Dementia appears to be associated with increased vessel resistance, the presence of spontaneous emboli and a reduced cerebrovascular response to increased/decreased environmental CO₂ using TCD methods. Blood flow velocities appear to decrease with age in late life. These patterns of TCD findings correspond to known structural vascular changes [16,62]. TCD techniques stand to make a valuable contribution to the understanding of underlying cerebrovascular contributions to age-related cognitive impairment and dementia. Further, TCD measures may assist in the development of novel therapeutic strategies addressing cerebral vasoreactivity or may be reliable methods to differentiate between dementia subtypes or predict clinical progression of cognitive decline [15,48,50].

H.A.D.K. is supported by an Australian National Health and Medical Research Council Training Fellowship (568890). This research was funded by the Brain Foundation, Australia (2012 Research Grant), and the Australian Association of Gerontology (2012 RM Gibson Scientific Research Award).