Phase-Contrast MRI Indices can Reect Intracranial Compliance Deterioration Induced by Intra-Abdominal/ Thoracic Hypertension

Identifying elevated intracranial pressure (ICP) and decreased intracranial compliance (ICC) is imperative for optimizing patient management in neurocritical care settings. Intra-abdominal hypertension (IAH) and intrathoracic hypertension (ITH) is common in trauma patients, which affects homeostasis of ICP/ICC. Knowledge of this effects is little and monitoring this effect is dicult. In the current study, we examined whether the indices generated from 2D cine phase contrast MRI (2D cine PC-MRI) could reect ICC/ICP alterations induced by elevated IAH/ITH during VM. Methods A total of 50 healthy young volunteers participated in this study (male: female = 24:26), and took a 2D cine PC-MRI during normal breath and VM respectively. Cross-section area (CSA) of dominant IJV and ipsilateral ICA, the maximum blood ow (F max ), minimum blood ow (F min ), mean blood ow (MBF), pulsatility index (PI), arteriovenous delay (AVD) and time to peak of arterial pulse (TTP) were gauged from images or calculated from the blood ow curves generated from 2D cine PC-MRI. These indices have potential clinical application in ICC/ICP monitoring in patients who was speculated with an IAH or ITH. The present work investigated the use of 2D cine PC-MRI as a means of assessment of ICC/ICP dynamic in young healthy subjects in a situation of elevated IAP/ITP modeled by VM. In this study, we demonstrate that elevated IAP/ITP causes ICC/ICP changes during VM, which can be revealed using 2D cine PC-MRI. Larger CSA IJV indicated an elevation of back pressure of IJV. Compared with baseline, smaller AVD indicates a decreased ICC or increased ICP, while larger TTP indicates a decreased ICC or increased ICP. These indices have a potential use in non-invasive ICC/ICP monitoring in patients with elevated IAP/ITP. study. Here, the the changes of these will be discussed. AVD, of PWV. AVD ICA IJV AVD in patients with hydrocephalus (NPH) [43] late-onset idiopathic aqueductal stenosis (LIAS) [44] Multiple sclerosis (MS) [45] NPH after CSF [43]


Introduction
The homeostasis of intracranial pressure (ICP) is of utmost importance for maintaining normal brain function. The content of the closed cranial vault is composed by the brain, venous-and arterial blood, and the cerebrospinal uid (CSF), all of which are virtually incompressible. Any expansion of one of these components must be compensated by a displacement of volume from one or more of the others [1] . This compensational capacity is de ned as intracranial compliance (ICC) [2] . Once the compliance runs out, there will be an exponential rise in ICP [2] . Thus, monitoring ICC dynamics would get insight into the subsequent ICP dynamics.
However, ICP/ICC is not only in uenced by intracranial contents, as ICP has been shown to have a complex relationship with both intra-abdominal pressure (IAP) and intrathoracic pressure (ITP). The ICP, IAP and ITP are closely interrelated in conditions of intra-abdominal hypertension (IAH), intrathoracic hypertension (ITH), abdominal compartment syndrome (ACS) and multicompartment syndrome (MCS), which were frequently encountered in patients with traumatic brain injury (TBI) and those complicating with multiple injury [3][4][5] . Impeded cerebral perfusion by intracranial hypertension (ICH) will lead to cerebral ischemia. Further uid administration to support cerebral perfusion or increasing ventilatory support to treat acute lung injury will further increase ICP, which can create a cycle that ultimately produces or aggravates MCS [6] . When it cannot be managed by maximal medical means, brain ischemia is the single most important factor increasing morbidity and mortality after TBI [6] . Therefore, optimization of patients with TBI should involves treatment of intracranial disorders and other organ systems that was cognizant of possible central nervous system effects [7] . IAH is a continuum from asymptomatic elevation of IAP to an immediately life-threatening situation, where dynamic evolution in both directions is possible. Therefore, it is di cult to identify triggers for interventions that may lead to complications or have adverse effects [8] . To date, no indicators for predicting the in uence of IAH/ITH on ICP have been reported. Timely detection of decreased ICC or raised ICP allows for early management, which reduces the risk of permanent brain damage, and improves patients outcome after brain injury [9] . However, direct monitoring ICC and ICP is di cult and risky, impeding it bene t for more patients. Several non-invasive methods had developed in last decades. Two-dimensional cine phase contrast MRI (2D cine PC-MRI) is one of these non-invasive methods. By gauging the parameters of cerebral hemodynamics and cerebrospinal uid dynamics, the 2D cine PC-MRI can estimate ICP/ICC qualitatively or quantitatively.
The Valsalva maneuver (VM), a natural response during coughing, defecation, and lifting heavy loads, is de ned as a forced exhalation against a closed glottis. The VM is commonly used to test cardiac functions or autonomic nervous system de ciencies [10] . Moreover, VM is also a well-established physiological model of ICH or IAH in basic and clinical research [11,12] . During VM state, increased IAP/ITP causes an increase in back pressure in the jugular veins (IJV), impeding venous blood drainage through IJV and leading to an increase of intracranial venous volume [13][14][15] . According to the Monroe-Kellie doctrine, a decrease of ICC or an elevation of ICP can be modeled [16,17] .
We conducted this study, aimed to investigate the effects of elevated IAP/ITP (modelled by VM) on ICC/ICP changes, and the potential indices that can be used in ICC/ICP monitoring in patients, using PC-MRI.

Page 3/10
The protocol was approved by the Ethics Committee of the General Hospital of the Chinese People's Liberation Army (Registry NO. S2015-014-02), and conformed to standards set by the Declaration of Helsinki.
A group of healthy young subjects were recruited for this study. Participants were informed of the potential risks and experimental procedures, and informed written consent was obtained. All procedures and protocols were approved by the ethics committee of the General Hospital of PLA and performed in accordance with the Declaration of Helsinki. Participants abstained from alcohol, caffeine, and strenuous exercise for at least 24 h before the MRI scan.

MRI Data Acquisition
All MRI examinations were performed on a 3.0T scanner (GE Discovery MR 750, Waukesha, WI) using an 8-channel head coil. Maximum intensity projection (MIP) of 2D-PC-MRV and T2 weighted images were used as localizer (Figure1). Velocity (encoding) sensitization was set at 50 cm/s, at which the internal carotid arteries (ICA) and internal jugular vein (IJV) can be imaged simultaneously. The plane that was perpendicular to bilateral ICA and IJV, passing through the level of odontoid process root of the second cervical vertebra, were selected for 2D cine PC-MRI acquisition (Figure1). The MRI parameters of 2D cine PC-MRI sequence were as follows: TR=7.0ms, TE=3.6ms, slice thickness=4mm, trigger time=19ms, acquisition matrix =224×224, ip angle =20°. Velocity (encoding) sensitization was set at 70 to 90 cm/s. Retrospective cardiac gating was selected, so that the 60 frames covered the entire cardiac circle (CC) were acquired. Each participant underwent the 2D cine PC-MRI sequence twice, once during normal breath (NB), and once during VM, using the same parameters and at the same plane. Acquisition time for once 2D cine PC-MRI was approximately 45s and 38s during NB and VM respectively, depending on the heart rate of individual. Total MRI scan was accomplished approximately in 10 mins. Before MRI scan, all participants practiced VM under the researcher's instructions. The sub-maximal VM (approximately 60% of one's maximal VM according to personal feeling) was adopted for 2D cine PC-MRI scan, for safety. Respiratory motion was monitored by placing bellows around the thoracoabdominal region, and PC-MRI scan started at the beginning of strain (phase a of VM ) manually (Figure1).

Measurement and indices calculation
All 2D cine PC-MRI data processing operations were performed with commercial workstations (GE, ADW 4.6) using Cardiac VX_1.1.0 software. Regions of interest (ROIs) was manually drawn around the boundary of the vessels of interest (bilateral ICA and IJV) on the amplitude map, referring to phase map when necessary, in the frame with maximum cross-sectional area (CSA), and the vessels were tracked automatically in all time frames. Blood ow was measured in ICA and IJV simultaneously. The blood ow curves of the CC were calculated by the 60 frames ( Figure 2). Then, the curve of IJV were reversed. The maximum ow (F max , ml/s), minimum ow (F min , ml/s) and mean blood ow (F mean , ml/s) and the CSA (mm 2 ) throughout the CC were generated against either the ow curve or image. According to the CSA of IJV during NB, the IJV with larger CSA and ipsilateral ICA were de ned as the dominant side [18] . The arteriovenous delay (AVD) were measured as the time delay between the F max of IJV (F max_IJV ) and F max of ipsilateral ICA (F max_ICA ). The time to peak of bilateral ICA ow curve (TTP ICA ) was the time took of the arterial blood ow from F min_ICA to F max_ICA . Both the absolute and normalized (divided by the time of CC, R TTP_ICA , R AVD ) TTP ICA and AVD were used for statistic calculation, to eliminate the in uence of heart rate ( gure 2). The pulsatility index (PI) were calculated according to previous reported method [19] , based on blood ow:  p=0.0329, p=0.038). From the blood ow curves, the peak of ow curve of IJV shifted to left, and the wave of blood ow in ICA broadened and the peak shifted to right (Table 4, gure3).

Key results
The present work investigated the use of 2D cine PC-MRI as a means of assessment of ICC/ICP dynamic in young healthy subjects in a situation of elevated IAP/ITP modeled by VM. In this study, we demonstrate that elevated IAP/ITP causes ICC/ICP changes during VM, which can be revealed using 2D cine PC-MRI.
Larger CSA IJV indicated an elevation of back pressure of IJV. Compared with baseline, smaller AVD indicates a decreased ICC or increased ICP, while larger TTP indicates a decreased ICC or increased ICP. These indices have a potential use in non-invasive ICC/ICP monitoring in patients with elevated IAP/ITP.

Interpretation the results
It had been universally recognized that the VM is a model of IAP and ITP elevation. In the present study, the dilation of IJV is and indicator of CVP elevation, which is caused by the increased IAP and ITP. Although most of the indices changed during VM, the PI was the candidate for measurement ICP in most previous studies. Despite AVD and TTP were reported only in very few studies, these two indices played a powerful role re ecting the VM induced ICC/ICP changes in the present study. Here, the interpretation of the changes of these indices will be discussed.
Interpretation of the decreased PI PI describes the pulsatility of blood ow waveform. Evidence suggests that either increased [19] or decreased [20] PI are harmful to the brain. During straining (VM phase ), central venous pressure (CVP) increased, cardiac output (CO) decreased and blood ow of middle cerebral artery dropped signi cantly [21] . Our result demonstrate the F max dropped more than F mean , while the F min remained constant, so the PI decreased. Dropped F mean indicated dropped cerebral perfusion pressure (CPP), that is consistent with previous study. However, although PI is commonly considered an indicator of CPP as opposed to ICP, there exists a complex relationship between PI and multiple hemodynamic variables including CPP, arterial pressure pulse amplitude, cerebrovascular resistance (CVR), arterial compliance, and heart rate [22] . Under normal conditions, PI is predicted to increase linearly with ICP [23,24] . While in a recent study, ICP estimated using PI method (ICP=a×PI + b) of ve in eight individuals were reported a trend of decrement during VM state [25] . Their results are consistent with the present study. A recent study also reported that increased ICP attenuates the pulsatility of cerebral venous out ow [26] .
Theoretically, venous out ow is a passive component, the intracranial pulse wave is predominately determined by the arterial wave [1] . Under conditions of the cardiac insu ciency and CVP increased, which is prevalent in critical ill patient and also in VM, venous pressure in uences the intracranial wave more [27,28] . During VM, the increased CVP impedes venous out ow, which in turn lead to an increment of cerebral blood volume and a "stiffer" brain. both the "stiffer" brain and the insu ciency cardiac function led to the CBF decrement and in particular the maximum blood ow. Thus, the pulsating component of the ICA and the IJV decreased.

Interpretation of the shortened AVD
Monroe-Kellie doctrine states that, the volumetric interplay between intracranial components (arterial blood, brain, venous blood and CSF) maintains ICP homeostasis [29] . In a cardiac circle, heart contraction induces a pulsatile artery blood entering the closed cranial cavity, which causes brain remolding and a transient perturbation of ICP homeostasis [30,31] . The arterial expansion must be compensated by venting CSF into the spinal canal and venous out ow [1,32] . A minimal time delay between arterial in ow and venous out ow from the cranial cavity has constantly been observed in previous studies [32][33][34][35][36] ,which was measured as AVD [37] .
Pulse wave velocity (PWV) or the time the pulse wave takes to travel the length of the artery was used to re ecting the compliance of the arterial tree [38,39] .
PWV has been considered the gold standard for measuring arterial stiffness [38,40,41] , higher PWV was independently associated with greater target organ damage [42] . Similarly, ICC can be estimated by measuring the PWV that traverses the intracranial space. In a certain distance, AVD, the time the pulse wave takes to travel from the arterial to the venous side of the cerebral circulation, is the reciprocal of PWV. AVD gauged from ICA to IJV at the skull base is a re ection of the compliance of the intracranial space, which includes arteries, subarachnoid space, brain parenchyma and veins. Smaller AVD have been reported in patients with normal pressure hydrocephalus (NPH) [43] , late-onset idiopathic aqueductal stenosis (LIAS) [44] and Multiple sclerosis (MS) [45] . What's more, larger AVD have been reported in patients with NPH after a CSF drainage than pre-drainage [43] .
There were few studies devoted to AVD measuring can be referred. In a study, AVD measured at the level of intervertebral disc of cervical-2 to cervical-3 using the same method, a smaller AVD (72±24 ms, representing 8±2% of the CC duration vs. 106 ) were reported in 18 (9 female, 9female) healthy young volunteers [46] . In another study, a larger AVD (17%±7% of the CC) in 19 (3 female, 16 female) healthy young volunteers were reported. This difference might be contributed to the different temporal resolution we used (32 frames per CC vs. 60 frames per CC we used). More accurate measure was expected by adopting a higher temporal resolution. In several other studies, AVD was measured as the time delay between the center of arterial pulse and the center of venous pulse, and the AVD was measured at much higher plane than the present study, so the results can't be compared directly [37,43] .
Under the VM state, forced expiration against a closed glottis increased IAP and ITP, damping IJV out ow and venting CSF upwards motion into the cranial cavity [47] , in turn, leading to an immediate stiffening of the brain and a rise of ICP (or a decrease of ICC) [48] . Shear waves propagate more quickly through a stiffer material [49] . Smaller AVD during VM than NB re ects the ICC decline during VM.
Interpretation of the prolonged TTP A recent study using invasive monitoring methods simultaneously recorded ICP and CBF waveform in non-human primate, reported that a higher ICP led to a broadening and rightward shift of the ICP and CBF pulse wave [50] , although the strength of this effect differed between subjects. In patients with NPH underwent CSF drainage, the arterial pulse became thinner, and peaked earlier than pre-drainage [43] . These results are consistent with the prolonged TTP during VM reported in our work. The TTP re ects the resistance of blood ow. Elevated IAP/ITP caused a decreased ICC and increased ICP, and in turn, the increment resistance of arterial perfusion, this has been reported in both animal model of IAH [51] and in man during VM [52] . In the present study, the smaller CSA ICA and decreased F mean also implied a restrictive effect of sympathetic activity during VM, which also led to an increase of vascular resistance.

Limitations
The strength of our study is fundamentally limited by that VM can only partially model IAH/ITH. (1) VM elevates IAP/ITP for a relatively short time, while IAH/ITH usually rises for hours to even for days in clinical scenarios; (2) All the subjects enrolled in this work were normal, that the respiration and circulation is different from patients with multiple trauma, in whom IAH/ITH often occurs after excessive uid therapy. Validation of these indices in clinical scenarios is needed. The second limitation is that mouth pressure, regarded as a surrogate for ITP, was not measured during VM. At a given increased ITP, diverse proportion of transmission to ICP was reported in previous studies [53] . Thus, the dynamic trends comparing to baselinerather than the absolute value of these indices are valuable when applicate this method [54] . The safety of a repeated MRI examination, the di culty of intracranial monitoring and the emergence and development of portable MRI [55][56][57] , suggest an oncoming MRI monitoring.

Conclusions
Using 2D cine PC-MRI and modelled by VM, this study investigated the in uence of elevated IAP/ITP on ICP/ICC. The results showed that the AVD between ICA and IJV, and time to peak of ICA pulse-wave have a potential role in monitoring the patients with IAH/ITH. The smaller AVD and larger TTP indicate a trend of ICC decrease and ICP rise.

Declarations Ethical Approval and Consent to participate
The protocol was approved by the Ethics Committee of the General Hospital of the Chinese People's Liberation Army (Registry NO. S2015-014-02), and conformed to standards set by the Declaration of Helsinki. All of the participants were informed of the potential risks and experimental procedures, and informed written consent was obtained.

Consent for publication
Not applicable.

Availability of supporting data
Data are the property of the authors and can be obtained by contacting the Principal Investigator: Dr. Jie F; e-mail: 13920449779@163.com

Competing interests
The authors declare that they have no competing interests.

Funding
This study was partially funded by 3 grants awarded to Lin Ma (81741115, and 16CXZ014 and 16BJZ11).
Authors' contributions J F and X Y performed data analysis and wrote the rst draft of the manuscript. L M conceived the study, J F, X Y, W L and J L contributed to data collection and analysis, and edited the manuscript.   Figure 2 Comparation of indices generated during NB and VM. During VM, the PI_ICA and PI_IJV decreased, AVD and RAVD decreased and the TTP_ICA increased signi cantly. ****, p<0.0001; * p<0.05. Figure 3