Discordant results from global versus local cerebral blood flow measurements in response to rhythmic handgrip exercise

Background The cerebral blood flow (CBF) response to exercise can be measured in a larger vessel such as the middle cerebral artery (MCA; e.g. by transcranial Doppler; TCD) and at tissue level by e.g. arterial spin labeling MRI (ASL). The reported changes in total CBF between artery flow and tissue flow approaches are, however, inconsistent [1]. The aim of the current study was to evaluate TCD and ASL measurements in response to the same protocol of rhythmic handgrip exercise in the same subjects, to determine into what degree large vessel blood flow measurements conform to local perfusion measures. As a second aim, the assumption of constant vessel cross‐sectional area (CSA) during handgrip exercise was assessed by high resolution MRI. Methods On 2 different sessions separated by 1 week, heart rate (HR) and end‐tidal CO 2 (P ET CO 2 ) were recorded in 11 subjects (5 female) during rhythmic handgrip. Three 5‐minutes runs of rhythmic handgripping at 60% of maximal voluntary contraction were alternated with 5 minutes of rest. During the first session regional CBF velocity (CBF v ) was measured in the MCA by TCD and during the second session tissue perfusion (CBF t ) was determined with ASL (3T Philips MRI, 17 slices). The CSA of the contralateral MCA was assessed with the same exercise protocol in a separate cohort of subjects (n=11, 5 female) using direct 7 Tesla MRI observations (following [2]). Results In response to rhythmic handgrip exercise, an increase in regional CBF v (5±6%, p<0.01) and motor area CBF t (10±6%, p<0.01) was demonstrated, while whole brain gray matter CBF t did not change (1±8% p=0.75). During both sessions, the increase in HR was similar (11±6% vs. 15±8%, p=0.81) whereas P ET CO 2 did not change. No differences in absolute resting values for HR (p=0.09) and mean arterial pressure (p=0.27) were observed. Resting PetCO 2 was slightly higher (2.3mmHg, p<0.01) during the MRI session. High resolution MR imaging demonstrated a decrease in the MCA CSA (−4.0±0.6%; p<0.001) in response to handgrip, together with an increase in HR (11±1%, p<0.001) whereas P ET CO 2 remained constant (p=0.72). Discussion The 10% CBF t increase as detected with ASL was restricted to the motor cortex, which represents only 2% of the total MCA flow‐territorial volume. When extrapolated to the total flow‐change within the MCA flow territory, this would represent only a 0.2% increase, as also confirmed by the non‐significant change of whole brain gray matter flow (1±8%). However, our TCD measurements indicated a 5% CBF v increase of the entire flow territory. These measures of CBF were therefore inconsistent and differed by a factor 25. The present findings could be explained by the observed vasoconstriction of the MCA, but also other physiological phenomena, such as arteriovenous shunts, or a breakdown of other assumptions of the applied technologies, such as flow profile changes might be involved. A limitation is that the two sub‐studies were performed in different subjects. Conclusion This study demonstrates discordant results from global (TCD) and local (ASL) cerebral blood flow measurements in response to rhythmic handgrip exercise. The observed reduction in MCA CSA implicates that the blood flow velocity changes measured by TCD overestimate the effects of mild exercise on CBF. Further pinpointing the nature of these discordant results will help interpretation of TCD and ASL studies concerning regulation of CBF. Support or Funding Information This study is supported by the Rembrandt Institute of Cardiovascular Science