Transcranial Doppler pulsatility index is not a reliable indicator of intracranial pressure in children with severe traumatic brain injury

Background: The TCD-derived PI has been associated with ICP in adult studies but has not been well investigated in children. We examined the relationship between PI and ICP and CPP in children with severe TBI. Methods: Data were prospectively collected from consecutive TCD studies in children with severe TBI undergoing ICP monitoring. Ipsilateral ICP and CPP values were examined with Spearman correlation coefficient (mean values and raw observations), with a GEE, and as binary values (1 and 20 mm Hg, respectively). Results: Thirty-four children underwent 275 TCD studies. There was a weak relationship between mean values of ICP and PI (P = .04, r = 0.36), but not when raw observations (P = .54) or GEE (P = .23) were used. Pulsatility index was 0.76 when ICP was lower than 20 mm Hg and 0.86 when ICP was 20 mm Hg or higher. When PI was 1 or higher, ICP was lower than 20 mm Hg in 62.5% (25 of 40 studies), and when ICP was 20 mm Hg or higher, PI was lower than 1 in 75% (46 of 61 studies). The sensitivity and specificity of a PI threshold of 1 for examining the ICP threshold of 20 mm Hg were 25% and 88%, respectively. The relationship between CPP and PI was stronger (P = .001, r -0.41), but there were too few observations below 50 mm Hg to examine PI at this threshold. Conclusion: The absolute value of the PI is not a reliable noninvasive indicator of ICP in children with severe TBI. Further study is required to examine the relationship between PI and a CPP threshold of 50 mm Hg. (C) 2009 Elsevier Inc. All rights reserved