Is it time for an autoregulation-oriented therapy in head-injured patients?

Optimization of cerebral perfusion pressure (CPP) is a cornerstone of the management of head-injured patients in intensive care. Unfortunately, there is still no consensus on the best strategies to manipulate CPP. Which should we modify first: blood pressure or intracranial pressure (ICP)? Which is the most appropriate drug? What are the pressure thresholds? Recent guidelines suggest that “CPP <50 mm Hg should be avoided, and that minimally invasive, efficient, and accurate methods of determining and following the relationships between CPP and autoregulation and between CPP and ischemia in individual patients are needed” (1). For the last 20 yrs, the Cambridge University group has been working on techniques to measure cerebral autoregulation in real time and has proposed many indexes retrospectively related with clinical and prognostic data (2–6). In the current issue of Critical Care Medicine, the retrospective, single- center study by Aries et al (7) adds new information for a bedside practical identification of an “optimal” value of CPP (CPPopt), which is measured as the CPP corresponding to the best cerebral vasoreactivity (8). The main result has been to develop an algorithm for the computerized, automated, and continuous updating of CPPopt, derived from a time window recording of 4 hrs. In their study, vasoreactivity has beencalculated as a linear correlation coefficient between spontaneous fluctuations of mean arterial pressure and ICP (PRx). Results are based on some theoretical considerations: Firstly, PRx is an index of vasoreactivity, which means that variations of ICP induced by CPP changes should be related to cerebral blood volume variations. In fact, if ICP is unchanged, cerebral blood volume should also remain unchanged. This is not always true, as the same authors state. In case of decompressive craniectomy, it is likely that spontaneous fluctuations of CPP do not induce ICP increase, even in the presence of significant raises in cerebral blood volume. The same can happen when cerebrospinal drainage is present, or cerebral compliance is high. Secondly, according to the authors, the best PRx should correspond to CPPopt. From a theoretical point of view, it is reasonable to consider that the best CPP for individual patients matches with the best cerebral vasoreactivity, but prospective validation studies are still lacking. In particular, we do not know if patients with high vasoreactivity present the best metabolic and perfusion indexes. For example, hypocapnia increases PRx, but may be dangerous for head-injured patients and may represent an important confounding factor that needs to be taken into account (9). An important step in the knowledge of this phenomenon is represented by Jaeger et al (10), who have found that brain tissue oxygen pressure increases according to CPP, only when measured CPP was below or equal to CPPopt. When measured CPP was higher than CPPopt, brain tissue oxygen pressure did not change, becoming independent from any increase in CPP. Even if brain tissue oxygen pressure is just a surrogate of cerebral blood flow, this could suggest that driving CPP in excess of CPPopt does not improve cerebral perfusion, at least in the area where the brain tissue oxygen pressure probe is inserted. Furthermore, we do not know if drugs such as mannitol, vasopressors, or other variables, such as hyperthermia, hypothermia, and fluctuations in sedation could impact on PRx and eventually on CPPopt extrapolation (11). In addition, even if a correlation between PRx and prognosis was reported by several studies, it was only retrospectively investigated (4–6). We do not know if optimizing CPP could improve prognosis, or if it simply reflects such a derangement of physiological parameters that is associated with poor outcome. Beside these theoretical limitations, there are important practical problems due to technical acquisition of arterial blood pressure and ICP and prospective artifact exclusion. Furthermore, independently from the quality of the pressure signal, in a significant number of cases a correlation between arterial blood pressure and ICP was lacking, consequently PRx was not available. In addition, technical difficulties in identifying CPPopt with an automated system are reported by the authors. The purpose of this study is ambitious and fascinating because it suggests an autoregulation-oriented strategy to identify individualized threshold of CPPopt, which correlates with patients’ prognosis. Such an approach has been already highlighted on head-injured patients by Howells et al (12). In an interesting article, they retrospectively compared the effects of ICP- and CPPoriented therapy in patients with continuous monitoring of cerebrovascular pressure reactivity. They found that CPP-oriented therapy was superior to ICP-oriented therapy when cerebrovascular reactivity was normal, but it was the reverse when cerebrovascular reactivity was impaired. The authors estimated that the correct CPP- or ICP-directed treatment could have, on average, increased the probability of a favorable outcome from 45% to 64%. This hypothesis is intriguing and needs a prospective validation. So far, no prospective randomized study has been done, and therefore, no definitive evidence supports the use of this technology in general practice. We look forward to an autoregulationoriented prospective randomized multicenter study in the near future.