The question evaluated for this systematic review was as follows: What is the effect of BS on CBF and CA in animals and humans?

BS was defined as an EEG pattern characterized by alternating periods of isoelectricity and active oscillations or “spikes” (2). Note, for the present study, “isoelectricity” was considered synonymous with BS as they both indicate a state of profound cerebral inactivity. BS could be induced by anaesthesia or by brain injury.

The measures of CBF that were included in the search were the following: measuring blood flow in the superior sagittal sinus, Xenon133 clearance, MS techniques (radiolabelled, florescent, or coloured), laser Doppler flowmetry, thermal diffusion flowmetry, positron emission tomography (PET) imaging, perfusion weighted magnetic resonance imaging (PW-MRI), arterial spin labeling magnetic resonance imaging (ASL-MRI), functional magnetic resonance imaging (fMRI) and blood-oxygen-level-dependent (BOLD), perfusion computed tomography (CTP), xenon-enhanced perfusion computed tomography (Xe-CT), single-photon emission computed tomography (SPECT), near infrared spectroscopy (NIRS), and transcranial doppler (TCD). The primary outcome of interest was the quantifiable effect that the induction of BS has on these measures of CBF. Animal studies included all studies performed in non-human animals including but not limited to rats, canines, swine, and non-human primates. Human studies included adults as well as pediatric populations.

Both continuous and intermittent methods of measures of CA were included in the search, a complete list of which can be found in Supplementary Appendix S1A.

Inclusion criteria for this study were as follows: BS measured by EEG in some form (visualized, bispectral index, power analysis); animal subjects, human participants, original full-length papers, use of some form of CBF measurement including the above, and/or a measurement of CA. Exclusion criteria included: non-English language and non-original studies.

MEDLINE, BIOSIS, Cochrane library, EMBASE, Global Health, and SCOPUS were all searched from their inceptions to August 2022 using relevant search terms. An example of the used search strategy can be found in Supplementary Appendix S1B. Search results were then compiled and deduplicated before being filtered for inclusion.

A two-step review of all articles returned by our search was conducted using two reviewers (A.Z.S. and L.F.). In the first filtering phase, reviewers independently screened all titles and abstracts for possible inclusion. Next, full texts of articles that passed the first filtering phase were assessed to confirm that they met all inclusion criteria. Any discrepancies between the two reviewers were resolved by the senior author (F.A.Z.). Finally, the reference lists for each included article were screened for any missed articles.

The following data was extracted from each of the final included articles and is presented in Supplementary Tables S1, S2: Study subject, experimental conditions, number of subjects, method of BS, method of BS determination, the measure of CBF, study results and conclusions, and study limitations.

Due to the heterogenous nature of the study designs and results of the relevant literature, no formal meta-analysis was conducted.

Given the overall goal of this review, to perform a comprehensive systematic review of the literature, a formal bias assessment was not conducted.

Of the 42 studies that examined animals, six used propofol (Artru et al., 1992; Ramani et al., 1992; Werner et al., 1992; Joshi et al., 2006; Wang et al., 2011; Liu et al., 2013a), 13 used barbiturates (thiopental) (Michenfelder, 1974; Kassell et al., 1980; Gronert et al., 1981; Milde et al., 1985; Nemoto et al., 1996a; Nemoto et al., 1996b; Klementavicius et al., 1996; Zarchin et al., 1998; Schmid-Elsaesser et al., 1999; Westermaier et al., 2000; Mäkiranta et al., 2002; Joshi et al., 2005; Hungerhuber et al., 2006), 18 used inhalational anaesthetics (isoflurane) (Newberg et al., 1983; Maekawa et al., 1986; Scheller et al., 1986; Baughman et al., 1989; Lutz et al., 1990; Zornow et al., 1990; Roald et al., 1991; Kochs et al., 1993; Golanov and Reis, 1995; Liu et al., 2011; Walter et al., 2013; Sutin et al., 2014; Choi et al., 2016; Benveniste et al., 2017; Yang et al., 2018; Zhang et al., 2019; Berndt et al., 2021; Sirmpilatze et al., 2022), four used a combination of methods (Harper and MacKenzie, 1977; Warner et al., 1989; Young et al., 1997; Joshi et al., 2004), and one induced BS through hypotension (Werner et al., 1995).

In terms of the subjects that were used in the studies, 12 used different species of dogs (beagle, mongrel) (Michenfelder, 1974; Kassell et al., 1980; Gronert et al., 1981; Newberg et al., 1983; Milde et al., 1985; Lutz et al., 1990; Zornow et al., 1990; Roald et al., 1991; Artru et al., 1992; Werner et al., 1992; Kochs et al., 1993; Werner et al., 1995), 18 used different species of rodents (Sprague Dawley rats, wistar rats, gerbils) (Maekawa et al., 1986; Baughman et al., 1989; Warner et al., 1989; Golanov and Reis, 1995; Nemoto et al., 1996a; Nemoto et al., 1996b; Klementavicius et al., 1996; Young et al., 1997; Zarchin et al., 1998; Schmid-Elsaesser et al., 1999; Westermaier et al., 2000; Hungerhuber et al., 2006; Liu et al., 2011; Liu et al., 2013a; Sutin et al., 2014; Choi et al., 2016; Benveniste et al., 2017; Berndt et al., 2021), six used different species of rabbits (Scheller et al., 1986; Ramani et al., 1992; Joshi et al., 2004; Joshi et al., 2005; Joshi et al., 2006; Wang et al., 2011), two used different species of swine (Mäkiranta et al., 2002; Walter et al., 2013), three used different species of non-human primates (Harper and MacKenzie, 1977; Yang et al., 2018; Zhang et al., 2019), and one used both rodents and non-human primates (Sirmpilatze et al., 2022).

Of the 24 studies that examined humans, six used propofol (Matta et al., 1995a; Newman et al., 1995; Doyle and Matta, 1999; Ludbrook et al., 2002; Klein et al., 2011; Chaix et al., 2020), four used barbiturates (thiopental) (Bendtsen et al., 1985; Young et al., 1991; Hoffman et al., 1997; Connolly et al., 2015), six used inhalational anaesthetics (isoflurane) (Lam et al., 1995; Artru et al., 1997; Reinsfelt et al., 2003; Reinsfelt et al., 2011; Golkowski et al., 2017; Sirmpilatze et al., 2022), six used a combination of methods (Woodcock et al., 1987; Matta et al., 1995b; Matta et al., 1999; Zanatta et al., 2013; Akbik et al., 2020), and three looked at the effect of brain injury on BS (Du et al., 2014; Chalia et al., 2016; Kassab et al., 2021).

In terms of the participants that were used in the studies, five involved adults undergoing cardiac surgery (coronary bypass graft or valve) (Woodcock et al., 1987; Newman et al., 1995; Reinsfelt et al., 2003; Reinsfelt et al., 2011; Zanatta et al., 2013), five involved adults undergoing open neurosurgical procedures (Bendtsen et al., 1985; Artru et al., 1997; Hoffman et al., 1997; Doyle and Matta, 1999; Klein et al., 2011), six involved adults undergoing orthopedic and other surgeries (Matta et al., 1995a; Matta et al., 1995b; Lam et al., 1995; Matta and Lam, 1995; Matta et al., 1999; Ludbrook et al., 2002), one involved patients undergoing carotid endarterectomy (Young et al., 1991), one involved patients undergoing minimally invasive neuroradiological procedures (Chaix et al., 2020), five involved critically ill and/or brain injured patients (Johnston et al., 2003; Connolly et al., 2015; Sahinovic et al., 2018; Akbik et al., 2020; Kassab et al., 2021), one involved infant patients with hypoxic ischemic encephalopathy (Chalia et al., 2016), and two involved healthy adults (Golkowski et al., 2017; Sirmpilatze et al., 2022).

In this systematically conducted scoping review, we identified 66 studies, most of which involved animals. This should be expected as manipulation of CBF using anaesthetics poses significant risk to humans and can only be investigated in participants who are receiving anaesthetics for some other purpose. This is also why most human studies involved patients who were injured or receiving a therapeutic surgery.

By far, most of the studies involved small animal models (rodents and rabbits), which are not as applicable to humans and is a weakness of the literature. The largest sample sized used in the small-animal models was 61 (Maekawa et al., 1986) and involved Sprague-Dawley rats. The largest sample size seen in the large animal models was 16 and involved non-human primates, but only two of the primates achieved BS (Zhang et al., 2019).

Six animal and seven human studies investigated propofol. Propofol is a γ-aminobutyric acid (GABA) agonist with a rapid terminal half life that readily crosses the blood brain barrier (Sahinovic et al., 2018). The most common adverse effects are bradycardia, hypotension, and very rarely, the life-threatening propofol infusion syndrome. In both animals and humans, propofol consistently showed a decrease in CBF in most studies of up to 70% in animals and up to a 44% decrease in humans. There were only two animal studies amongst the studies that showed a change in CBF that examined CMR_O2 (Artru et al., 1992; Werner et al., 1992). One found that the CBF decreased to the same proportion as the CMR_O2 (Werner et al., 1992) and another found that the CBF decreased disproportionately much greater compared to the decrease in CMR_O2 (Artru et al., 1992). The former result would be expected if the decrease in CBF was purely due to flow-metabolism coupling whereas the latter would imply that propofol may have also had direct effects on the smooth muscle of the cerebral blood vessels themselves in addition to causing a reduction in CMR. Interestingly, in the human studies, while five studies demonstrated a significant decrease in CBF (Matta et al., 1995a; Newman et al., 1995; Doyle and Matta, 1999; Ludbrook et al., 2002; Klein et al., 2011; Chaix et al., 2020), all but one (Klein et al., 2011) showed no significant change in the AVD_O2, supporting the conclusion that propofol decreases CBF through mechanisms other than flow-metabolism coupling. In the one study that investigated patients with TBI, even though CBF was not measured, there was no significant effect found on ICP or AVD_O2 (Johnston et al., 2003). Interestingly, in one animal study, propofol reduced CBF by similar amounts under both normocapnia and hypocapnia (Artru et al., 1992) and in one human study, decreased CBF by similar amounts under both hypothermia and normothermia (Newman et al., 1995). Both hypothermia and hypercapnia are known to reduce CBF through vasoconstriction (Tameem and Krovvidi, 2013) and interestingly, it seems that when propofol is added, despite the fact that it reduces CBF on its own, the effect does not seem to be synergistic. Interestingly, it seems that patients who have cardiovascular risk factors have a higher reduction in CBF compared with those who have fewer risk factors (Chaix et al., 2020). In patients with hypertension, due to factors such as arterial remodelling and renin-angiotensin system, the CA curve is shifted to the right and cerebral ischemia can occur at higher MAPs compared to patients without hypertension (Tameem and Krovvidi, 2013; Machado et al., 2021).

Barbiturates have a similar mechanism of action to propofol but have longer half-lives and have higher risk of inducing hypotension (Shin et al., 2018). The number of animal studies that investigated barbiturates was comparable to the other anaesthetics; this contrasted with the human studies of which only four investigated barbiturates. Eleven of the animal students demonstrated that barbiturates had a significant reduction in CBF to a maximum of 80% (Michenfelder, 1974; Kassell et al., 1980; Gronert et al., 1981; Milde et al., 1985; Nemoto et al., 1996a; Nemoto et al., 1996b; Klementavicius et al., 1996; Zarchin et al., 1998; Schmid-Elsaesser et al., 1999; Westermaier et al., 2000; Hungerhuber et al., 2006). Six animal studies showed a decrease in CMR_O2 (Michenfelder, 1974; Kassell et al., 1980; Milde et al., 1985; Nemoto et al., 1996a; Nemoto et al., 1996b; Klementavicius et al., 1996) and out of those six studies, five showed a decrease proportional to the decrease in CBF (Michenfelder, 1974; Kassell et al., 1980; Nemoto et al., 1996a; Nemoto et al., 1996b; Klementavicius et al., 1996). One study even found a strong positive correlation between CBF and CMR_O2 (R = 0.77; p < 0.05) (Kassell et al., 1980). In another study in which CBF decreased disproportionately more than CMR_O2 (75% vs. 54%), the anaesthetic used to induce BS was etomidate and CBF plateaued before BS was even reached (Milde et al., 1985). As mentioned earlier, etomidate is not technically a barbiturate and is structurally unrelated to other anaesthetics (Williams et al., 2022). Etomidate has been shown to cause depression in certain parts of the brain and cause hyperexcitability in others such as the basal ganglia. However, in the one human study that used etomidate for anaesthesia, the CBF and CMR_O2 decreased proportionally by 40% (Hoffman et al., 1997). Therefore, it remains unclear that is the effect etomidate has on CBF is due to its decrease in CMR alone or also due to other mechanisms. Interestingly, the only animal study in which CMR_O2 was increased used pentobarbital (Gronert et al., 1981). This was presumably due to tolerance as there is evidence to suggest that tolerance develops much quicker with pentobarbital compared to the other barbarities (Saunders et al., 1990). There was a significant lack of human studies that investigated barbiturates; out of the four, one used etomidate (Hoffman et al., 1997) and one only had a sample size of 2 (Connolly et al., 2015). The lack of studies investigating barbiturates on humans may be due to its unpopularity as an anaesthetic given their tendency to cause hypotension (Shin et al., 2018).

It has been well-described that inhalational anaesthetics produce deep sedation by reducing activation in the thalamus and midbrain reticular formations (Campagna et al., 2003). However, while there is a reduction in brain activation in several cortical areas, activation of the visual, motor, and subcortical regions seems to be unaffected. While inhalational anaesthetics primarily target the GABA receptor, they have effects on a plethora of other channels including glycine receptors, nicotinic acetylcholine receptors, and glutamate receptors (Campagna et al., 2003). There were a total of 18 animal and 6 human studies that used inhalational anaesthetics. The animal literature was extremely variable in terms of the effect that inhalational anaesthetics have on CBF and only four showed a decrease in CBF (Baughman et al., 1989; Zornow et al., 1990; Walter et al., 2013; Berndt et al., 2021). This may have been because, as previously mentioned, inhalational anaesthetics have different effects on the activity of different parts of the brain. In addition, the studies used a variety of concentrations of isoflurane, ranging from 1%–3%. In one study, at a concentration of 1.0 minimum alveolar concentration (MAC), there was no change in CBF but at 1.5 MAC the auditory cortex saw a 4.9% significant decrease in CBF (p < 0.05), the visual cortex did not see any significant change, and the extrapyramidal system (140%–170%) and limbic system (87%–120%) saw a significant increase, and finally, with 2.0 MAC, all regions saw a significant increase in CBF (Maekawa et al., 1986). Further, in one fMRI study, there was more widespread fluctuation in BOLD signal and significant correlation with cortical sensory areas and the thalamus during BS (Sirmpilatze et al., 2022) and in another a lower amplitude of the BOLD signal and significant correlation of the BOLD signal with BS in frontal, parietal, and temporal lobes and in basal ganglia (Golkowski et al., 2017). While CMR_glu decreased across the brain, it did not decrease in the hippocampus or corpus callosum at any concentrations (Maekawa et al., 1986). Several of the animal inhalational anaesthetic studies (Golanov and Reis, 1995; Liu et al., 2011; Sutin et al., 2014; Zhang et al., 2019; Berndt et al., 2021) demonstrated strong correlation between EEG bursts/spikes and spikes in CBF, lending support to the mechanisms behind flow-metabolism coupling in the control of CBF. Out of the six human studies that used inhalational anaesthetics, four showed a consistent decrease in CBF or CBF velocity (Roald et al., 1991; Kochs et al., 1993; Liu et al., 2013b; Yang et al., 2018). While sevoflurane did reduce CBF velocity and CBF (Artru et al., 1997; Reinsfelt et al., 2011), but it did not show any significant reduction in CPP (Artru et al., 1997). This was in contrast to isoflurane which showed a 36% reduction in CPP in the same study (Artru et al., 1997). None of the human studies examined the effect of the anaesthetics on CMR_O2 or AVD_O2 and therefore, many questions remain unanswered regarding the effect that isoflurane and other inhalational drugs have on CBF and to what extent the effect is mediated by a change in CMR.

In terms of the studies that investigated multiple anaesthetics, there was evidence to suggest that when inhalational anaesthetics are added to propofol, there is an increase in CBF velocity compared to the use of propofol alone. This may be because inhalational anaesthetics increase CBF in certain parts of the brain, thereby decreasing the effect that propofol has. Further, inhalational anaesthetics affect a variety of receptor sites including glutamate receptors and by doing so may nullify the effect of propofol (e.g., inhibition of inhibition). Compared to inhalational anaesthetics, barbiturates seem to have a higher effect on reducing CBF, which was seen in an animal (Warner et al., 1989) as well as a human study (Woodcock et al., 1987). This may be due in part to the higher likelihood of barbiturates to cause hypotension but may also be because inhalational anaesthetics have different effects depending on the region in the brain (Campagna et al., 2003; Golkowski et al., 2017; Sirmpilatze et al., 2022).

Overall, there was a significant lack of discussion of autoregulation in the studies that were examined in the present review, as will be discussed in the limitations section. Previous work has shown that, at lower doses, propofol, barbiturates, and inhalational anaesthetics do not affect CBF and autoregulation (Veselis et al., 2005; Farag et al., 2017; Slupe and Kirsch, 2018) at low doses but at higher doses CBF and autoregulation are impaired (Slupe and Kirsch, 2018). It was unclear why higher doses of anaesthetics cause this change. Given that most of the evidence pointed toward BS causing a reduction in CBF that was disproportionately greater than the reduction in MAP (Michenfelder, 1974; Kassell et al., 1980; Gronert et al., 1981; Milde et al., 1985; Artru et al., 1992; Ramani et al., 1992; Werner et al., 1992; Matta et al., 1995b; Newman et al., 1995; Nemoto et al., 1996a; Doyle and Matta, 1999; Westermaier et al., 2000; Ludbrook et al., 2002; Joshi et al., 2006; Chaix et al., 2020), it can possibly be inferred that BS has at least some effect on autoregulation impairment. While there was some evidence that CBF changes in proportion to CMR_O2 (Michenfelder, 1974; Kassell et al., 1980; Milde et al., 1985; Werner et al., 1992; Nemoto et al., 1996a; Klementavicius et al., 1996), there was also evidence that CBF changed disproportionately (Milde et al., 1985; Baughman et al., 1989; Artru et al., 1992), making it challenging to know what effect exactly BS has on autoregulation. It is also possible that different depths of sedation have a significant effect on animal and human cerebral autoregulation. For example, previous work looked at the relationship between objectively measured depth of sedation (through the bispectrality index) and cerebrovascular reactivity (surrogate measure of cerebral autoregulation) in patients with TBI (Froese et al., 2022a; Froese et al., 2022b). From this work it was seen in almost all patients that there was a depth of sedation that optimized cerebrovascular reactivity (achieve the most intact cerebral reactivity value), indicating that too much/little sedation results in non-optimal cerebral states.

In most of the studies, it was clear that BS decreases CBF, and this effect is most pronounced with propofol and barbiturates. However, it remains unclear why BS has this effect. CBF is believed to be controlled by four primary mechanisms: myogenic, endothelial, neurogenic, and flow-metabolism coupling (Tameem and Krovvidi, 2013). In the myogenic mechanisms, the smooth muscle cells themselves are responsible to modulating cerebral vessel diameter. The endothelial theory suggests that shear-stress induced by blood flow changes on the endothelial lining leading to induction of vasoactive mediators, such as nitric oxide or endothelin, precipitating changes in vascular tone. In the neurogenic mechanism, vasoconstriction is caused by norepinephrine and neuropeptide Y released from postganglionic sympathetic from the superior cervical ganglion. Finally, in flow-metabolic coupling, CBF is closely matched to the demands of the neuronal firing. Flow-metabolic coupling is thought to be mediated by astrocytes, a type of glial cell in the brain. While the process is not entirely understood, it is believed that astrocytes communicate the activity of neurons to blood vessels and coordinate the energy demand with oxygen and glucose supply (Venkat et al., 2016). Astrocytes may mediate the synthesis of vasoactive substances, thereby inducing vasodilation when neuronal activity increases. Reasonably, it is unlikely that the anaesthetics have much of an effect on the smooth muscle cells or on the sympathetic ganglion as much as they do on the neuronal activity. However, given that many studies did not investigate CMR_O2 and a significant few looked at CMR_glu, it is challenging to know if this is indeed the mechanism. One of the main confounders in the human studies was the maintenance of MAP. In some of the studies, MAP was maintained with pressors above a given level with pressors while in others, it was allowed to drop. First, if MAP is allowed to drop, it is difficult to state whether CBF dropped due to a drop in systemic hemodynamics or due to the BS induced by the anaesthetic. Interestingly, in previous papers our team has shown that, while pressors cause vasoconstriction when applied to cerebral blood vessels directly, there is no effect when applied systemically (Froese et al., 2020b; Froese et al., 2020c). Therefore, vasopressors do not appear to have a significant effect on CBF or autoregulation. Without MAP being maintained throughout the experiment, it is challenging to know if CBF dropped due to alteration of cerebral hemodynamics and therefore a myogenic mechanism or due to a flow-metabolic coupling method. Further, without estimating the amount of sedation that is given to the subjects, no firm conclusions can be drawn in relation to the effects that BS induced by anaesthetics has on CBF. More animal but especially more human studies are needed to clarify how and why BS causes a decrease in CBF.

It remains unclear what effect inhalational anaesthetics have on CBF as there were mixed results from the studies. These anaesthetics are especially interesting as they seem to not have as much hemodynamic depression (Campagna et al., 2003) as other agents and are especially useful for disease processes such as refractory status epilepticus (Mirsattari et al., 2004). Further, if inhalational anaesthetics are used in combination with other anaesthetics in the neurocritical care unit, could the profiles of both agents be used in a synergistic fashion? For example, adding isoflurane to propofol in a patient in whom it would be beneficial to reduce the CMR but not the CBF as much. There still exists a gap in the literature regarding inhalational anaesthetics as well as the use of a combination of anaesthetics in patients. There was a significant lack of studies that investigated the effect of BS on CBF in patients with critical neurological illness or traumatic brain injury. Given that these patients may have altered autoregulation (Armstead, 2016), it would be pertinent to see how these patients respond to the different anaesthetics.

Finally, most of the studies that were examined in the present review did not measure the effect of BS on autoregulation. From the results of the present review, we can conclude that BS leads to a decrease in CBF but given that autoregulation links CBF and CMR_O2 by flow-metabolism coupling, the drop in CBF may be an appropriate response of autoregulation (Tameem and Krovvidi, 2013). There are two overarching methods of measuring autoregulation: a “static” method and a “dynamic” method (Tiecks et al., 1995). The older, static method involves measuring CBF at a constant baseline MAP and constant CBF. After manipulation of the MAP is undertaken, another measure of CBF is taken and if this measure changes significantly, there is an impairment in CA (Tiecks et al., 1995). The major disadvantage with this method is that there is no assessment of the time in which the change in CVR occurs. The dynamic method involves using a thigh blood pressure cuff (Obrist et al., 1975); the cuff is inflated and rapidly released which causes rapid drops in MAP. During this process, MAP and CBF velocity, measured using TCD, and the latency of the change in CVR can be measured. In using both methods with BS, there exists a confound. One possible method of inferring a loss of autoregulation is if CBF decreases disproportionately more than CMR_O2 as they should otherwise be tightly coupled. More studies are needed to measure the effect of BS on CBF and CMR_O2 using both static and dynamic methods.

One of the main limitations to the present review was its breadth. By design, the review’s purpose was to examine the effect of BS on overall cerebral physiology, hence, we were not selective about the methods used to measure CBF or to induce BS. The advantage to this was that we were able to assess the effect of BS in a wide variety of scenarios and in different species. However, by including a vast array of methodologies, we compromised on accuracy. For example, we included studies that used TCD in the present review even though TCD measures CBF velocity which is not completely analogous to CBF and under conditions in which the vessel diameter is variable, can under or overestimate CBF (Coverdale et al., 2014). In addition, we included studies that measured CBF using the BOLD signal, which only measures changes in oxygenated blood (Logothetis, 2002). Nevertheless, inclusion of these studies allowed us to examine the effect of BS on cerebral physiology in a variety of experimental conditions.

One limitation of the present review is that we did not compare the different techniques used to measure CBF. A number of different techniques were used in the studies mentioned in this review and some are known to be more accurate than others. However, the primary goal of the review was to take the techniques as a whole to determine what the consensus was on the effect of BS and CBF, and despite the variety of methods used, there was agreement amongst most of the studies. In future work, it may be beneficial to examine the different techniques used and their accuracy for measuring CBF.

Another limitation was that we focused on the quantitative effect that BS has on CBF but did not look at as much on the clinical implications of this in animals, in healthy controls or in patients with brain injury. For example, if certain anaesthetics are superior at reducing CMR_O2, they might be more neuroprotective or if others have a high effect on CBF, they may not be the optimal choice for patients at risk for brain ischemia. In future work, we hope to look at the clinical significance of different anaesthetics and how this relates to their specific action on CBF.

Finally, in the present review, we were not able to conduct a meta-analysis of the data due to extremely heterogenous populations and data sets. In future work, it would be pertinent to narrow our search to studies in which there are more homogenous populations (e.g., patients with status epilepticus) and determine the effect size that BS has on CBF or even on clinical outcomes.

Future work should use in vivo methods of monitoring CBF such as NIRS in neurologically ill patients to determine the effect of BS in this population. The NIRS technique uses the infrared light, which can penetrate into deep tissues, and be absorbed by oxygenated and deoxygenated hemoglobin (Chen et al., 2020). The technique can determine the regional concentration of oxyhemoglobin and deoxyhemoglobin and thereby, the oxygenation of the tissue. In other words, NIRS is a cost-effective, portable, and efficient means of measuring neurovascular coupling. The advantage of being able to monitor cerebral physiology continuously and non-invasively in critically ill patients make it a favored technique for further investigating the questions left unanswered by the present review. A previous scoping review of 47 studies by our team showed evidence that continuous-wave NIRS metrics were linearly correlated with changes in CBF (Gomez et al., 2022). Cerebral oxygenation can be obtained from NIRS, and by correlating this value to MAP, one can obtain the cerebral oximetry index (COx) which is a measure of CA (Chen et al., 2020). When COx is close to zero or negative, autoregulation is being maintained but when the value is closer to one, autoregulation is impaired. Using NIRS, EEG, and BS, the effect of the above-discussed anaesthetics could be investigated in patients admitted to the neurological ICU in real time. Other evidence has shown that NIRS can predict epileptic spikes even before the spikes are seen on EEG (Chen et al., 2020). In patients with status epilepticus, the technique could be used to investigate CBF, autoregulation, and metabolism in this critically ill group of patients to determine how best to use BS to improve morbidity and mortality. Concurrent work by our group aims to answer these questions using continuous in-vivo methods of neuromonitoring such as NIRS.

In this systematic review of the effect of BS on CBF and autoregulation in animals and humans, we identified 67 studies. Most of the literature involved animal models, which largely involved small animals. In the human studies, most were conducted on patients undergoing surgery or who were critically ill, as expected. There was significant evidence to suggest that propofol and barbiturates decrease CBF in both animals and humans. However, the evidence regarding inhalational anaesthetics, comparative effects, and clinical implications is lacking. More large animal models as well as clinically relevant human studies are needed to further elucidate the mechanisms behind BS and CBF and what implications this has on the future of care in the neurological intensive care unit.