Comparison of Transcranial Doppler and Ultrasound-Tagged Near Infrared Spectroscopy for Measuring Relative Changes in Cerebral Blood Flow in Human Subjects

BACKGROUND: Currently, no reliable method exists for continuous, noninvasive measurements of absolute cerebral blood flow (CBF). We sought to determine how changes measured by ultrasound-tagged near-infrared spectroscopy (UT-NIRS) compare with changes in CBF as measured by transcranial Doppler (TCD) in healthy volunteers during profound hypocapnia and hypercapnia. METHODS: Ten healthy volunteers were monitored with a combination of TCD, UT-NIRS (c-FLOW, Ornim Medical), as well as heart rate, blood pressure, end-tidal Pco(2) (PEtco(2)), end-tidal O-2, and inspired O-2. Inspired CO2 and minute ventilation were controlled to achieve 5 stable plateau goals of Etco(2) at 1520, 25- 30, 35-40, 45-50, and 55-60 mm Hg, for a total of 7 measurements per subject. CBF was assessed at a steady state, with the TCD designated as the reference standard. The primary analysis was a linear mixed-effect model of TCD and UT-NIRS flow with PEtco(2), which accounts for repeated measures. Receiver operating characteristic curves were determined for detection of changes in CBF. RESULTS: Hyperventilation (nadir PEtco(2) 17.1 +/- 2.4) resulted in significantly decreased mean flow velocity of the middle cerebral artery from baseline (to 79% +/- 22%), but not a consistent decrease in UT- NIRS cerebral flow velocity index (n = 10; 101% +/- 6% of baseline). Hypercapnia ( peak PEtco(2) 59.3 +/- 3.3) resulted in a significant increase from baseline in both mean flow velocity of the middle cerebral artery (153% +/- 25%) and UT- NIRS (119% +/- 11%). Comparing slopes versus PEtco(2) as a percent of baseline for the TCD (1.7% [ 1.5%- 2%]) and UT- NIRS (0.4% [0.3%- 0.5%]) shows that the UT- NIRS slope is significantly flatter, P <.0001. Area under the receiver operating characteristic curve was significantly higher for the TCD than for UT- NIRS, 0.97 (95% confidence interval, 0.92-0.99) versus 0.75 ( 95% confidence interval,0.66-0.82). CONCLUSIONS: Our data indicate that UT-NIRS cerebral flow velocity index detects changes in CBF only during hypercarbia but not hypocarbia in healthy subjects and with much less sensitivity than TCD. Additional refinement and validation are needed before widespread clinical utilization of UT-NIRS.