Effect of controlled blood pressure increase on cerebral blood flow velocity and oxygenation in patients with subarachnoid haemorrhage

Patients with aneurysmal subarachnoid haemorrhage (SAH) might have impaired cerebral autoregulation, that is, CBF – and thereby oxygen delivery – passively increase with an increase in CPP. This physiological study aimed to investigate the cerebral haemodynamic effects of controlled blood pressure increase in the early phase after SAH before any signs of delayed cerebral ischaemia (DCI) occurred.


| INTRODUCTION
Of the patients surviving spontaneous aneurysmal subarachnoid haemorrhage (SAH), approximately 50% have an unfavourable neurological outcome. 1,2 The pathogenesis and pathophysiology behind the poor outcomes are complex and related to the extent of initial brain injury, but also complications such as re-bleeding, hydrocephalus, and delayed cerebral ishaemia. [3][4][5] Both in the very early phase (0-3 days after the bleeding) and in the slightly later phase where delayed cerebral ischaemia may occur (from 3 days and onwards), impaired regulation of cerebral blood flow (CBF) may play a role. 6 In one observational study of patients with poor-grade SAH undergoing neuromonitoring, a lower cerebral perfusion pressure (CPP) in the early phase was associated with a higher risk of cerebral metabolic crisis, characterised by a low level of glucose and a high lactate-pyruvate ratio in cerebral microdialysate, and of brain tissue hypoxia; in turn, brain tissue hypoxia and metabolic crisis were independently associated with a poor functional outcome at three months. 7 The most likely explanation is that patients with poor-grade SAH have impaired cerebral autoregulation, meaning that an increased CPP increases CBF as well as glucose and oxygen delivery to the brain. 8 However, the effects of direct manipulation of perfusion pressure on CBF and glucose/oxygen delivery was not assessed in these patients.
In the present study, we aimed to investigate how a brief course of controlled blood pressure increase affects CBF measured by MCAv (our primary outcome) and cerebral oxygenation measured by PbtO2 (the secondary outcome)within 5 days of ictus, that is, before the potential occurrence of delayed cerebral ischaemia (DCI). 6 We measured middle cerebral artery flow velocity (MCAv) by transcranial Doppler (TCD) as a surrogate for CBF, local brain tissue oxygen tension (PbtO 2 ), and cerebral microdialysate markers of glucose, oxidative metabolism, and glycerol. We hypothesised that a noradrenalineinduced increase in mean arterial blood pressure (MAP) would increase CPP, MCAv, PbtO 2 , and microdialysate glucose, and reduce microdialysate lactate-pyruvate (LP)-ratio, glutamate, and glycerol. 14 May 2019). The study was preregistered at clinicaltrials.gov (NCT03987139; 14 June 2019) before the inclusion of the first participant. The primary outcome was MCAv, which was chosen as a surrogate of CBF (see Appendix S1, Protocol deviation). The exploratory outcomes comprised PbtO 2 , ICP, arterial oxygen tension (PaO 2 ), arterial carbon dioxide tension (PaCO 2 ), arterial oxygen content (CaO 2 ), global oxygen delivery (DO 2 ), and four microdialysis markers (glucose, LP-ratio, glutamate, and glycerol) (Table S1).

| Study population
Adult patients admitted with SAH to the Neurointensive Care Unit, Rigshospitalet, Denmark, were screened for inclusion. The inclusion criteria were (1) SAH; (2) placement of an external ventricular drain (EVD); (3) possible to conduct the study within 5 days of ictus; and (4) patient or relatives understand oral and written Danish. The exclusion criteria were (1) aneurysm not secured, either because treatment failed or a conservative approach was adopted; (2) pupils fully dilated and unresponsive to light on admission; (3) brain herniation before inclusion or expected death within 48 h; and (4) other diseases or conditions associated with impaired autoregulation (e.g., DCI, ischaemic stroke, TBI, bacterial meningitis or sepsis within a year; or diabetes mellitus with organ manifestations). DCI was defined by Vergouwen et al. 6 as a clinical diagnosisseldom present before day 5diagnosed by a drop in GCS by 2 or more point without any other explanation. Informed consent was obtained from patients and/or next of kin depending on the capacity of the patient. The clinical management of patients with SAH at Rigshospitalet is described in Appendix S1.

| Measurements
Baseline measurements were recorded during a period of 20 min where the patient was clinically stable, and where only the most necessary changes in sedatives, analgesics, vasopressors, and ventilator settings were allowed. MCAv was recorded unilaterally through the temporal window by TCD using a 2 MHz probe (Multi-Dop T, DWL, Singen, Germany) fastened by a LAM rack (DWL, Atlanta, GA). MAP was measured invasively through a radial artery catheter with the transducer placed at the external acoustic meatus, and ICP was measured with a Codman Microsensor ICP Transducer (Integra Life-Sciences, Princeton, NJ) or an ICP sensor attached to an external ventricular drain (Spiegelberg, Hamburg, Germany). Arterial blood gases were sampled at the end of each period for measurement or calculation of PaO 2 , PaCO 2 , CaO 2 , and DO 2 (For calculations, see Appendix S1). For participants with PbtO 2 and microdialysis (n = 17), PbtO 2 was recorded continuously (Licox, Integra Lifesciences, Princeton, NJ). Cerebral microdialysis was conducted with a sampling rate of 0.3 μL/min through a catheter with a 20 kDa semipermeable membrane (M Dialysis AB, Stockholm, Sweden) and hourly analysis; measurements were adjusted so that one vial of microdialysate was removed immediately before initiation of controlled blood pressure increase, and one vial was removed 1 h later.

| Intervention
After the baseline period, controlled blood pressure increase was induced by increasing the dosage of noradrenaline. We aimed at increasing the MAP by a maximum of 30 mmHg and to an absolute level of no more than 130 mmHg. 9 This increase is considered safe and was based on the clinical recommendations for induced hypertension after DCI. The period of controlled blood pressure increase started after a steady state of the desired MAP was achieved for at least 20 minutes, following which noradrenaline was tapered until the MAP at baseline was achieved.

| Sample size calculation
The calculated sample size was based on MCAv as the primary outcome. The paired t-test sample size estimation resulted in 34 participants and was based on a pragmatically chosen minimum relevant difference (MIREDIF) in MCAv of 6 cm/s, a standard deviation (SD) of 12 cm/s, 10 a two-tailed α of 0.05 and a power of 0.8 (1 À β).

| Data collection and calculations
Demographic data were collected and handled using REDCap samples. An alpha below or equal to 0.05 was considered significant for the primary outcome, while the p values for the exploratory outcomes were presented after Benjamini-Hochberg correction. 12 Associations between different measurements were explored by scatter plot and calculation of the corresponding Pearson's correlation coefficient.
Arterial blood gas values remained unchanged after the controlled blood pressure increase ( Figure S3). The controlled blood pressureinducedchange in PbtO 2 did not correlate to changes in CPP, MCAv, CaO 2 , or DO 2 ( Figure 2). Intracerebral microdialysis measurements of glucose, LP-ratio, glutamate, and glycerol showed no changes upon the controlled blood pressure increase ( Figure S6).

| DISCUSSION
In the present study, we investigated the effects of a controlled noradrenaline-induced blood pressure increase in 36 patients with SAH and found that although a an increased CPP was associated with an increase in PbtO 2 , neither MCAv nor microdialysis-based indices of cerebral ischaemia were affected.
Previous studies of patients with cerebral tumours have reported a CBF increase during vasopressor infusion as measured by magnetic resonance imaging and positron emission tomography. 13 MCAv still remained unchanged. An unchanged MCAv could be associated with an increased PbtO 2 (indicating increased local blood flow) only assuming that the MCAv did not reflect local blood flow, the local oxygen demand (i.e., CMRO 2 ) decreased, or the perfusion distribution was changed. The potential limitations of MCAv as a marker of global CBF have been discussed above. A reduction in local oxygen demand appears unlikely, since noradrenaline infusion does not reduce the oxygen demand in brain tissuethat is, CMRO 2in healthy volunteers 16  optimal microvascular oxygen delivery in the brain. Increased transittime heterogeneity is associated with inhomogeneous oxygen extraction and, in effect, reduced tissue oxygenation even in the face of a CBF increase. 17 Animal studies indicate that cerebral capillary transittime heterogeneity is pathologically increased after SAH 18,19 and that it may be reduced or near-normalised through a vasopressor-induced increase in CPP. 20 Furthermore, in patients with SAH, it has been suggested that increased cerebral capillary transit-time heterogeneity may contribute to DCI. 19 On the other hand, another alpha-agonist, phenylephrine, increased rather than decreased cerebral capillary transit-time heterogeneity in patients with cerebral tumours. 21  investigate the effect of a controlled blood pressure increase during episodes of DCI, during which blood flow is commonly assumed to be reduced. 23 In poor-grade SAH patients, lower perfusion pressures are associated with a higher risk of brain tissue hypoxia and metabolic crisis, which in turn are associated with poor metabolic outcomes. 7 However, no randomised trial has been conducted to address whether active maintenance of a high perfusion pressure after the aneurysm has been secured, but in the pre-DCI phase, reduces the risk of cerebral metabolic crisis and a poor functional outcome in SAH patients.

| Strengths/limitations
This study assessed the effect of a direct change in CPP on CBF, rather than spontaneously occurring changes; spontaneous changes are often associated with or even caused by variations in other physiological variables such as arterial carbon dioxide or oxygen tension, body temperature, or metabolism, all of which may in themselves affect CBF. As an important limitation, the increase in MAP was less than planned, which increased the risk of type II errors; however, we do not believe that the absence of increase in MCAv was due to a type II error, as the data bordered on a decrease. A substantial increase in the noradrenaline infusion rate or intracranial pressure increase was among the factors limiting our ability to reach the predefined thresholds, and in this physiological study we chose to err on the side of caution. Moreover, this study was limited by not having a direct measure of the dynamic cerebral autoregulation. Unfortunately, in studies run in parallel to the present study, we found the preplanned measure of cerebral autoregulationmean flow indexto be unreliable and invalid. [24][25][26][27] Finally, as mentioned above we measured MCAv as a surrogate of global CBF.

| CONCLUSIONS
In this physiological study of patients in the early stage after SAH, a brief increase in mean arterial pressure induced by noradrenaline infusion showed no increase in MCAv but was associated with an increase in PbtO 2 . Our results suggest that autoregulation may not be impaired in these patients, in which case other mechanisms than a CBF increase may mediate the increase in brain oxygenation. Alternatively, a CBF increase did occur that, in turn, increased cerebral oxygenation, but was not detected by TCD. Consent to participate: Informed consent was obtained by the next-of-kin.