Critical Thresholds for Cerebrovascular Reactivity: Fact or Fiction?

Abstract

To the Editor, We read with interest the recent publication by Sorrentino et al. [1] entitled: ‘‘Critical Thresholds for Cerebrovascular Reactivity after Traumatic Brain Injury’’. The study is based on the hypothesis that bedside real-time calculation of the pressure-reactivity index (PRx) allows a continuous estimation of cerebral pressure autoregulation [1–3]. From measurements of hydrostatic pressure alone the authors define threshold levels for survival as well as favorable outcome in patients with severe traumatic brain lesions. Pressure autoregulation is present in many tissues but it is most pronounced in the brain and the kidney. The main ‘‘physiological purpose’’ of pressure autoregulation is probably to keep intracapillary hydrostatic pressure relatively constant [4]. Changes in precapillary resistance (R) underlying cerebral pressure autoregulation are calculated from cerebral perfusion pressure (CPP) divided by cerebral blood flow (F): R = (Pa ICP)/F [5]. Accordingly, to describe cerebral vasoreactivity it is necessary to assess not only arterial blood pressure (Pa) and intracranial pressure (ICP) but also F. The authors behind the PRx method claim that cerebrovascular pressure reactivity may be estimated by observing the response of ICP to changes in Pa. PRx is in this study determined by calculating the correlation coefficient between 30 consecutive data points resulting from the time averaging of ICP and Pa signals with a width of the moving-average window of 8 s [1]. However, in our opinion it is obvious from the equation above that cerebrovascular resistance cannot be calculated unless cerebral blood flow is also assessed. The method of PRx is based on the assumption that the observed change in Pa is the primary and independent cause of the variations in the other three monitored variables. In the validity study of the PRx method Pa was increased by i.v. vasopressor infusion and F was measured utilizing PET-technique [3]. In this defined, experimental situation it is possible that the increase in Pa is the primary and independent cause of the variations in the other variables. Under clinical conditions this is, however, often not the case. During neurocritical care fluctuations in pain, sedation, stress level, local neuronal excitation, etc., will cause changes in cerebral energy metabolism resulting in variations in precapillary vasoconstriction and cerebral blood flow that are not a function of simultaneously occurring changes in Pa. All these sources of error are neglected when PRx is interpreted as a measure of cerebral pressure autoregulation. As the Cambridge group has also suggested that PRx is correlated to cerebral energy metabolism [6] it seems unlikely that this index might be used for continuous estimation pressure autoregulation as well. In their analyses of prognostic threshold levels the authors use mean values for the entire monitoring period for each parameter: ICP, CPP, and PRx. In the study, the monitoring periods vary between 6 h and 20 days. The authors comment on this fact in the section of limitations. However, in our opinion the consequences are more serious than mentioned by the authors and may invalidate the threshold levels arrived at. In patients with brain trauma mortality is usually caused by an increase in ICP and patients who die within the first few days of admission usually have a high ICP. However, patients who initially may have equally high ICP (in this C.-H. Nordstrom (&) T. H. Nielsen Department of Neurosurgery, Odense University Hospital, Odense, Denmark e-mail: carl-henrik.nordstrom@med.lu.se