A bright forecast for autoregulation monitoring?

Cerebral autoregulation is a dynamic physiological process that sustains stable cerebral blood flow in the face of varying cerebral perfusion pressure. Continuous autoregulation monitoring following acute brain injury has been proposed as a novel treatment target against which to optimise cerebral perfusion pressure and thereby maintain cerebral blood flow to avoid secondary brain injury. This proposed treatment paradigm involves manipulating arterial blood pressure (and hence cerebral perfusion pressure) with vasopressor and inotropic medication, guided by a continuously monitored index of autoregulation (Aries et al. 2012).

In a recent article in The Journal of Physiology Liu et al. (2018) describe the validation of a new iteration of this paradigm, which addresses limitations of existing autoregulation indices, and may therefore deliver a more reliable clinical treatment target. In brief, it has been observed that a theoretically optimal perfusion pressure (CPP‐opt) can be derived from a surrogate of autoregulation, the pressure reactivity index (PRx). PRx, derived from intracranial pressure and arterial blood pressure, reaches a nadir consistent with an “optimal” level of autoregulation (CPP‐opt), where blood pressure changes are not associated with, or are negatively associated with, intracranial pressure changes (Aries et al. 2012). In observational studies patients managed close to CPP‐opt have improved outcome following traumatic brain injury and it has been postulated that this might reflect reduction of hypoperfusion‐related cerebral hypoxia–ischaemia and hyperaemia/oedema (Liu et al. 2018). Considerable interest exists to examine this technique in a randomized clinical trial but no ideal monitor of autoregulation exists that can be measured continuously at the bedside and fully encompass the complexity of the dynamic physiology of autoregulation. Liu et al. (2018) address this challenge by validating a new index based on the PRx that is optimised to track dynamic changes.

The PRx index is derived from spontaneous, low frequency oscillations in intracranial pressure and arterial blood pressure by Pearson correlation of 5 min signal data epochs (Aries et al. 2012). Intracranial pressure is thought to indirectly reflect cerebral blood volume and hence the radius of resistance arteries. When autoregulation is impaired or perfusion pressure is below the lower limit of autoregulation, slow spontaneous oscillations of blood pressure are entrained into the intracranial pressure signal. Thus a high PRx reflects impaired autoregulation, and low or negative PRx reflects intact autoregulation. CPP‐opt is derived by tracking PRx across time and evaluating the distribution of the relationship between PRx and perfusion pressure (Aries et al. 2012). Where blood pressure oscillations are optimally buffered, there is no relationship, or a negative relationship, between blood pressure and intracranial pressure. One critical limiting factor is the dynamic nature of autoregulation physiology and hence oscillations varying both in time and frequency; thus the PRx has temporal limitations given the 5 min windowing, and cannot optimally track physiological processes that may vary dynamically in frequency, such as cerebral autoregulation. This may negatively impact CPP‐opt evaluation.

Liu et al. (2018) validate a new technique in an animal model using an advanced wavelet signal processing methodology, which has a key advantage: it allows the continuous analysis of both frequency and time domain. They demonstrate that this new wavelet index, termed wPRx, agrees with PRx in two animal models where cerebral perfusion pressure is manipulated experimentally. Recently wavelets have been increasingly used to analyse autoregulation (Highton et al. 2015; Chalak & Zhang, 2017). This facilitates examination of dynamic signal properties and in the case of evaluating CPP‐opt has the potential to offer greater CPP‐opt accuracy. Two wavelet properties of the oscillations are of interest, and employed by Liu et al. (2018): coherence and phase. Coherence represents similarity in power (without respect to direction) whilst phase reflects the direction of the relationship. Intact autoregulation buffers changes in perfusion pressure and as such coherence is reduced and phase unrelated to, or opposite to perfusion pressure. Larger changes in perfusion pressure may evoke a brisk autoregulation response which results in a negative relationship. Thus both coherence and phase can be used as parameters to quantify autoregulation (Chalak & Zhang, 2017). Where coherence is low phase measurement is unreliable, therefore coherence can also be used (as in Liu et al. 2018) to discriminate signal content which is likely to reflect autoregulation rather than noise. The wPRx is based on wavelet semblance, a measure of phase first described for forecasting geophysical time series and subsequently employed in autoregulation analysis (Highton et al. 2015) because of similar behaviour to PRx.

Crucially the wPRx has been robustly validated against PRx in a controlled animal model, and as such Liu et al. (2018) is a key foundation for future research using wavelet semblance. The utility of wPRx has now subsequently been confirmed in a large dataset of patients with traumatic brain injury, offering a wPRx derived CPP‐opt which is more closely associated with functional outcome (Liu et al. 2017). However, further work is required to identify if wPRx can be used to guide autoregulation oriented therapy, and furthermore if this improves metabolism or secondary injury. Wavelet semblance has also recently been used to compare blood pressure to mitochondrial redox oscillations (cytochrome c oxidase) in neonatal hypoxia–ischaemia (Mitra et al. 2017) which similarly is associated with magnetic resonance spectroscopy evidence of energy failure and impaired outcome. Wavelet techniques and the wPRx may therefore further enhance our understanding of haemodynamic and metabolic changes following brain injury and are a promising treatment target for a future randomized clinical trial of autoregulation oriented therapy.

Additional information

Competing interests

None declared.