Send to

Choose Destination
Stroke. 2020 Feb;51(2):468-474. doi: 10.1161/STROKEAHA.119.027829. Epub 2019 Dec 30.

Confounding of Cerebral Blood Flow Velocity by Blood Pressure During Breath Holding or Hyperventilation in Transient Ischemic Attack or Stroke.

Author information

From the Department of Clinical Neuroscience, Wolfson Centre for Prevention of Stroke and Dementia, University of Oxford, United Kingdom (A.J.S.W., S.M., L.L., P.M.R.).
Headache and Neurosonology Unit, Neurology Department, Università Campus Bio-Medico, Rome, Italy (M.P.).


Background and Purpose- Breath holding (BH) and hyperventilation are used to assess abnormal cerebrovascular reactivity, often in relation to severity of small vessel disease and risk of stroke with carotid stenosis, but responses may be confounded by blood pressure (BP) changes. We compared effects of BP and end-tidal carbon dioxide (etCO2) on middle cerebral artery mean flow velocity (MFV) in consecutive transient ischemic attack and minor stroke patients. Methods- In the population-based, prospective OXVASC (Oxford Vascular Study) phenotyped cohort, change in MFV on transcranial Doppler ultrasound (ΔMFV, DWL-DopplerBox), beat-to-beat BP (Finometer), and etCO2 was measured during 30 seconds of BH or hyperventilation. Two blinded reviewers independently assessed recording quality. Dependence of ΔMFV on ΔBP and ΔetCO2 was determined by general linear models, stratified by quartiles. Results- Four hundred eighty-eight of 602 (81%) patients with adequate bone windows had high-quality recordings, more often in younger participants (64.6 versus 68.7 years; P<0.01), whereas 426 had hyperventilation tests (70.7%). During BH, ΔMFV was correlated with a rise in mean blood pressure (MBP; r2=0.15, P<0.001) but not ΔCO2 (r2=0.002, P=0.32), except in patients with ΔMBP <10% (r2=0.13, P<0.001). In contrast during hyperventilation, the fall in MFV was similarly correlated with reduction in CO2 and reduction in MBP (ΔCO2: r2=0.13, P<0.001; ΔMBP: r2=0.12, P<0.001), with a slightly greater effect of ΔCO2 when ΔMBP was <10% (r2=0.15). Stratifying by quartile, MFV increased linearly during BH across quartiles of ΔMBP, with no increase with ΔetCO2. In contrast, during hyperventilation, MFV decreased linearly with ΔetCO2, independent of ΔMBP. Conclusions- In older patients with recent transient ischemic attack or minor stroke, cerebral blood flow responses to BH were confounded by BP changes but reflected etCO2 change during hyperventilation. Correct interpretation of cerebrovascular reactivity responses to etCO2, including in small vessel disease and carotid stenosis, requires concurrent BP measurement.


blood pressure; humans; ischemic attack, transient; leukoaraiosis; linear models; stroke

Supplemental Content

Full text links

Icon for Atypon
Loading ...
Support Center