Cerebral perfusion targets set according to cerebrovascular reactivity have been suggested and investigated as one of the strategies for individualised management in traumatic brain injury (TBI) patients admitted to critical care.1–3 While extensive retrospective research has been conducted over the past twenty years, the prospective evaluation of such strategies is at its infancy3. As a consequence, the physiological effect of targeting cerebral perfusion pressure (CPP) at the “optimal” cerebral perfusion pressure (CPPopt) tailored to optimise cerebrovascular reactivity, as a surrogate for cerebral autoregulation, is still unknown.
The hypothesis behind the autoregulation-guided management is that by optimising cerebral autoregulation, episodes of brain hypoperfusion or hyperaemia are avoided as the brain is maintained at a CPP where it can best maintain adequate cerebral blood flow in the face of systemic changes.4 This should translate in better clinical outcome by limiting the secondary injury.5,6 Therefore, targeting CPP at the level at which autoregulation is best preserved, should translate in achieving the best autoregulation possible. The index of autoregulation used for evaluating CPPopt is the pressure reactivity index (PRx) 7. It is logical to postulate that PRx should indicate a better-preserved vascular reactivity if CPP is at CPPopt levels.
To date, the only concluded prospective interventional trial on autoregulation-guided management in TBI patients is the ‘CPPopt Guided Therapy: Assessment of Target Effectiveness’ (COGiTATE) trial 8, which showed that targeting CPP at CPPopt is feasible and safe in TBI patients requiring intracranial pressure management (NCT02982122, www.clinicaltrials.gov). In the intervention group of the study, CPP targets were adjusted 4-hourly according to an automated estimation of CPPopt 9, assessed via customised ICM+ software at the bedside. In the control group, CPP was targeted between 60 and 70 mmHg according to the Brain Trauma Foundation guidelines. Although the trial was not powered for effectiveness, it was notable that the grand mean value of PRx was not significantly different between the control and intervention group of the COGiTATE trial (mean (SD) of 0.0331 (0.199) vs –0.0417 (0.231), p = 0.2, t-test). While this might seem unexpected, there are several reasons that could explain such result, beyond the power of the study. In the next paragraphs we will provide the physiological rationale we followed for investigating those reasons.
The relationship between CPP and PRx can be in general described as a U-shape curve (clearly discernible in some percentage of cases), constrained to PRx range [-1;1]. The optimum value of this PRx/CPP curve is identified by an x-optimum (CPPopt) and a y-optimum (PRxopt). The latter can correspond to different levels of PRx, defining the vertical location of the curve within the plot. The shape of such a curve can also differ based on the y-span of the curve within the physiological CPP limits (see figure 1). If PRxopt is markedly positive, the curve has a small autoregulatory-span and the vascular reactivity would be impaired at any CPP level. As a consequence, targeting CPPopt would not allow to move from a non-autoregulating to an autoregulating physiology, but simply to maintain the best possible reactivity for that period. On the contrary, in cases when PRxopt is negative, and the shape of the curve has a wide y-span (PRx ranges from negative to positive within the physiological CPP values), it is more plausible that targeting CPPopt would provide means for achieving a good vascular reactivity. A simple metric of deviation of CPP from CPPopt (deltaCPPopt) does not convey the difference between those two cases (see figure 1 for more details). Similarly, a simple comparison of the averages of PRx between patient groups would also very likely be heavily affected by those effects. In this paper we explore the behaviour of PRx at different granularity in the COGiTATE cohort, comparing control and intervention group. We hypothesize that in the intervention group PRx was closer to PRxopt values, indicating a more preserved reactivity, as opposed to the control group.
Further, the feasibility results of the COGiTATE study showed that the percentage of time with patients’ CPP concordant to the CPPopt based target value was on average 46.5% in the intervention group 8. It is possible that this creates a dilution effect on the difference in PRx. As a proof of concept, we investigated the difference in PRx values comparing periods with CPP within the autoregulation target versus periods with CPP outside the autoregulation target, for individual patients. We hypothesise that in the intervention group, patients experienced a better-preserved vascular reactivity when CPP was within the CPPopt based target.