Introduction

Vascular cognitive impairment (VCI) defines the wide spectrum of cognitive disorders caused by different types of cerebrovascular disease. Within this frame, the term VCI-no dementia (VCI-ND) describes those subjects whose cognitive decline is not associated with substantial functional impairment. Although VCI-ND is not necessarily predictive of a progression into overt dementia [1], it represents a higher risk to develop more severe forms of cognitive impairment either as major VCI or mixed degenerative and vascular dementia (VaD), especially after recurrent strokes [2].

Growing evidence correlates cerebral hypoperfusion to both cognitive decline and white matter lesions (WMLs) [3]. Previous studies have applied transcranial Doppler ultrasound (TCD) to explore the relationship between cerebral hemodynamics and brain lesions attributed to small vessel disease in cognitive disorders [3–6]. The rationale is that microangiopathy, demonstrated in both vascular and degenerative dementia, might lead to arteriolosclerotic processes, vasoconstriction, and vascular stiffness, thus resulting in decreased arterial diameter and cerebral blood flow [3,7,8]. In this scenario, TCD is a well-known and feasible tool to evaluate the cerebral hemodynamics, the arterial perfusion integrity, and the intracranial small vessel compliance [9–11].

Although TCD has been employed in patients with VaD [3,7,8,12,13], no comparable data are available in those with VCI-ND. The aim of this study was to assess indices of cerebral blood flow velocity in VCI-ND patients and to correlate TCD changes with neuropsychological scores and WML severity. We hypothesized that alterations in cerebral hemodynamic may affect cognition even in the absence of overt dementia.

Methods

Results

From a total sample of 413 participants that were originally screened, 103 were excluded due to an MMSE ≤ 24 or a loss of autonomy according to the ADL and/or IADL scores, 25 presented with insufficient/absent acoustic transtemporal windows bilaterally, MRI was contraindicated in 21 patients due to different reasons (metallic implants, pacemakers and implantable defibrillators, claustrophobia, etc.), and 6 refused to participate. Therefore, a total of 258 subjects satisfying the study criteria were finally enrolled and classified into two groups based on the results of the neuropsychological tests: 161 VCI-ND patients (Group A) and 97 patients with WMLs without any cognitive deficit (Group B). Demographic features, vascular risk factors, neuropsychological scores, neuroradiological findings, and TCD measures are reported in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Clinical-demographic characteristics, neuroradiological features, neuropsychological test scores, and TCD values of all participants.

https://doi.org/10.1371/journal.pone.0216162.t001

None of the participants had focal motor deficits; slight reflex asymmetry was present in 16 patients of Group A and 9 of Group B, whereas unsteadiness was detected in 6 and 2 patients, respectively. Group A patients were older (p< 0.001) and less educated than Group B (p = 0.006), whereas sex did not differ between these groups. No difference in vascular risk factors was found, except for coronaropathy that was more prevalent in Group A (p = 0.042). In addition, Group A patients showed more severe WMLs than Group B (p< 0.01).

The mean insonation depth of the M1 segment of the middle cerebral artery was 55 ± 5 mm. An adequate transtemporal window for TCD measurements was achieved in 90.3% of the study subjects. Compared to Group B, Group A exhibited lower values for MBFV and EDV (p< 0.01), but higher values for PI and RI (p< 0.01; see S1 Fig).

The Spearman correlation detected a significant association between all TCD parameters and Stroop T. In particular, PI, RI, and PSV were directly correlated to Stroop T (p< 0.01), whereas PI and RI were inversely correlated to MMSE (p< 0.05) and MoCA (p< 0.01). The values for MBFV and EDV were inversely correlated to Stroop T values (p< 0.01), whereas the EDV was directly correlated to the scores in the MoCA (p< 0.01) and the MMSE (p = 0.04; see S2 Fig).

Table 2 shows the regression models in Group A. PI was significantly and independently associated with Stroop T (p< 0.001) and WML severity (p< 0.001). Similarly, PI (p< 0.001), as well as RI (p< 0.001) and EDV (p = 0.001) were associated with Stroop T and WML severity, whereas PSV was associated with Stroop T (p = 0.028). A reduction in MBFV did not predict cognitive dysfunction or WML severity. The multivariate linear regression models performed in Group B did not produce any significant result. The analysis adjusted for age, education, and vascular risk factors did not show any influence of these factors on the aforementioned results. Moreover, even when regression has included the whole sample of subjects (Group A + B), the results did not differ in terms of lack of significant effect of age. This variable, indeed, did not reach the statistical significance needed as a confounding factor.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Backward linear regression: Significant predictors of TCD parameters in VCI-ND patients (Group A).

https://doi.org/10.1371/journal.pone.0216162.t002

Discussion

To the best of our knowledge, this is the first TCD study investigating cerebral hemodynamics in VCI-ND patients. While different studies investigating cerebral hemodynamics in Mild Cognitive Impairment (MCI) of degenerative cause have been published, no study evaluating TCD parameters in a large cohort of non-demented patients with cognitive decline of vascular origin (VCI-ND) has been available until now.

The main finding of this study is that specific measures of cerebral perfusion and vascular resistance were significantly associated with WMLs and executive performance in patients with VCI-ND. It is worth to highlight that the literature provides limited data on the hemodynamic changes in the preclinical stages of dementias. In particular, previous reports failed to demonstrate PI or MBFV differences between patients with MCI and healthy controls [27,28]. Conversely, PI was increased in both VaD [8,29] and Alzheimer’s disease (AD) with respect to healthy controls [12,27,29], although some reports found a higher PI in VaD than in AD [13,30].

In the present study, although the MBFV was lower in VCI-ND patients than in those without, it was not independently predictive of cognitive impairment, possibly reflecting only a partially altered cerebral hemodynamics at this stage. A reduced MBFV might be due to the hypoperfusion secondary to arterial sclerosis [31]. Compared to PI, however, MBFV seems to be a weaker TCD index for impaired cerebral hemodynamics in cognitively impaired subjects. Indeed, previous studies did not show a significant difference in MBFV between demented patients and controls [30,32]. Nevertheless, a recent meta-analysis concluded that MBFV is lower in both AD and VaD patients compared to controls, although disturbances in cerebral hemodynamics were more severe in VaD, suggesting a specific spectrum of hemodynamic changes in different types of dementia [7].

Overall, the TCD changes detected in VCI-ND revealed a global pattern of cerebral hypoperfusion and increased vascular resistance, likely due to microcirculation pathology and linked to small vessel and capillary damage. These findings have been extensively demonstrated in VaD [29] but also in AD [33] and are primarily related to lipohyalinosis processes, supporting the role of vascular contribution in neurodegeneration. Furthermore, the increase of PSV, RI, and PI that we observed in VCI-ND patients can be considered as the correlate of distal arterial constriction or increased downstream resistance. Indeed, the increase in PI indirectly indicates a rise of vascular peripheral resistances and might also be explained by the diastolic reduction in flow velocities and the cerebral vascular compliance decline [3,7,8]. Therefore, the combination of high vascular resistances and low perfusion state suggests a global vascular impairment, possibly starting from small vessels and then spreading to the larger arteries. In this regard, some authors proposed that PI represents a marker of small vessel disease [34]. Interestingly, the severity of hemodynamic alterations detected by TCD might also reflect the severity of WMLs. Accordingly, we found that increased PI values resulted independently from other factors in more extensive WMLs on brain MRI [5,34,35]. Similarly, PI was a predictor of both worse cognitive performance and brain lesion load.

The results of this study should also consider that the subcortical white matter is markedly vulnerable to hypoperfusion and subsequent ischemia because of poorly vascularization by perforating arteries [36]. Hypoperfusion might cause ischemic injury of white matter tracts and interruption of the subcortical circuits, thus resulting in various cognitive symptoms and, particularly, in executive dysfunction [37,38]. Hence, TCD flow pattern of the large arteries stiffness is also associated with WML burden.

A strength of the present study is the enrolment of a large and representative sample of patients with VCI-ND caused by subcortical ischemic vascular disease. Subjects were also stratified according to their WML load, which is a marker of small vessel disease progression. Additionally, the TCD findings observed may contribute to the further understanding of the process and progress of VCI and represent an additional tool to disentangle the mixed forms of cognitive decline, even in the early stages. Lastly, VaD can be prevented. Therefore, TCD may be added as a clinical routine examination to detect patients who can benefit from an intensive medical therapy. One of the limitations is the lack of healthy controls, although the difficulties in recruiting a comparable number of age-matched healthy subjects without any imaging evidence of subcortical ischemic vascular disease (commonly observed among elderly) and cognitive impairment in neuropsychological tests should be taken into account. Moreover, the recruitment was conducted in a single center, and no follow-up was carried out.

Conclusion

TCD provides useful indices of correlation between WMLs and executive dysfunction in VCI-ND. Since the cerebral hypoperfusion is regarded as both a risk factor and an aggravating component of cognitive decline in the elderly, TCD represents a valuable tool to screen subjects at risk of both vascular and mixed dementia. Future studies are needed to acquire a more comprehensive understanding of the complex relationships between hemodynamic changes, WMLs, and cognitive decline. The possibility to early detect any marker of transition into overt dementia will support clinicians towards a more careful diagnosis and management of this population at risk.

Supporting information

Acknowledgments

We would like to thank Editage (https://www.editage.com) for English language editing.