ABSTRACT. The peak amplitude of pulsatile cerebral electrical impedance (A&) was compared with simultaneous '33xenon clearance estimations of cerebral blood flow (CBF,) on 28 occasions in nine infants receiving assisted ventilation who had changes in Pacoz and thereby presumably in cerebral blood flow. Percentage changes from one measurement to the next in each infant were compared.Using linear regression the relationship was A& = 0.5 CBF, -1.5 with r = 0.67. The 95% confidence interval for the regression coefficient was 0.2-0.8 and the mean residual was 28%. Changes in cerebral blood flow in these clinical conditions were similarly detected by the two methods but A& underestimated the magnitude of the change in comparison with CBF, and its accuracy was insufficient to allow conclusions about the magnitude of small changes in cerebral blood flow in individual infants. (Pediatr Res 24: 461-464,1988) Abbreviations A&, peak amplitude of pulsatile cerebral electrical impedance CBF,, cerebral blood flow estimated from '33xenon clearance curve extrapolated to infinity TcPco2, partial pressure of skin carbon dioxide TcPoz, partial pressure of skin oxygen Cerebral blood flow is assumed to be a vital determinant of neurologic outcome in the sick newborn infant. Information on brain blood flow has been hampered by lack of suitable measurement techniques. '33Xenon clearance is a standard for accurate, quantitative measurement of cerebral blood flow but is limited by the time-averaged nature of the result (because it is necessary to calculate this from the clearance curve that must be collected over some time), the inability to make rapid serial measurements and the total number of studies being determined by the safe limits of ionising radiation exposure. The cerebral electrical impedance technique provides continuous information noninvasively, but the biophysical basis of the technique remains incompletely understood. It relies on the fact that the brain and blood offer a very different resistance to the passage of electric
46current. As blood volume increases in the head after cardiac systole, impedance falls. The cardiac-synchronous component of cerebral impedance is determined predominantly by the instantaneous intracranial blood volume (1). Attempts to correlate cerebral electrical impedance with cerebral blood flow have been made in dogs (1) and in adults (2, 3) with variable results. Modeling the thorax or limb as a cylinder allows reasonably accurate estimates of flow to be made (4-6), but extrapolation of the cylindrical model to the brain is inappropriate because of the complexity of cerebral blood flow. In the human newborn the technique of cerebral electrical impedance has been used to show a reduction in the impedance during tension pneumothorax when cerebral blood flow would be expected to have fallen and an increase in impedance after feeding when cerebral blood flow would be expected to have risen (7). It has also been compared with strain gauge plethysmographic assessment of cerebral blood flow and in...