Intracranial Pressure and Cerebral Hemodynamic Monitoring After Cardiac Arrest in Pediatric Pigs Using Contrast Ultrasound-Derived Parameters

Abstract

Objectives

We explore the correlation of contrast-enhanced ultrasound (CEUS) parameters to intracranial pressure (ICP) in a porcine experimental model of pediatric cardiac arrest.

Methods

Eleven pediatric pigs underwent electrically induced cardiac arrest followed by cardiopulmonary resuscitation. ICP was measured using intracranial bolt monitor and CEUS was monitored through a cranial window. Various CEUS parameters were monitored at baseline, immediately post return of spontaneous circulation (ROSC), 1 hour-post ROSC, and 3 hours post-ROSC.

Results

There was significant ICP correlation with wash-out slope assessed by CEUS time intensity curve analysis at immediate post-ROSC. At 3 hours post-ROSC there was also significant negative correlation between ICP and peak enhancement which may be due to the evolution of anoxic injury.

Conclusion

The use of CEUS in assessing disruption of cerebral hemodynamics and ICP post cardiac arrest will need future validation and comparison to other imaging modalities. The correlation between CEUS parameters and ICP may be due to the alterations in cerebral autoregulation that result from anoxic brain injury.

Animal models have shown the rapid decline of vital substrates at the time of cardiac arrest: there is depletion of brain tissue oxygen tension after 0 to 2 minutes of cardiac arrest in pigs,¹ and loss of 75 to 80% of adenosine triphosphate at approximately 4 minutes after global ischemia in canine brains.^2,3 Brain is the most vulnerable organ to ischemic injury and there is a complex autoregulatory mechanism to prevent and mitigate factors that contribute to ischemia.

Understanding this physiology and monitoring hemodynamic parameters are of paramount importance in clinical management of patients who suffer from cardiac arrest or hypoxic ischemic injury. Cerebral hemodynamic monitoring may allow clinicians to better titrate therapies in the post-arrest period to mitigate secondary neurologic injury. It is important to note, however, that post-return of spontaneous circulation (ROSC) neuromonitoring is rarely used in pediatric patients and a noninvasive modality for doing so may bridge this gap. Beyond cerebral hemodynamic monitoring, assessing for potential elevations in intracranial pressure (ICP) is critical to preventing devastating intracranial consequences such as hemorrhage or brain herniation. Thus far, the gold standard method of ICP measurement is by an invasive approach such as placement of intracranial monitoring device or external ventricular drain. Given the procedural risks involved, such as infection and catheter tract hemorrhage,^4–8 noninvasive monitoring may provide clinicians with a safer and more tolerable modality of neuromonitoring following cardiac arrest. Additionally, perfusion measurements that entail time and resource-intensive methods such as perfusion magnetic resonance imaging, or methods that expose subjects to radiation such as perfusion-computed tomography pose significant limitations in studying pathophysiology of hypoxic ischemic injury especially in children.

In the recent years, there have been developments in various methods of monitoring ICP and cerebral hemodynamics. These include transcranial Doppler (TCD), optic nerve sheath diameter measurement using ultrasound, and near infrared spectroscopy. TCD is applied at the acoustic window such as a fontanelle in infants as well as the temporal bone in children and adults. The ultrasound beam is emitted through these acoustic windows and a reflected beam from flowing erythrocytes allows for the calculation of macrovascular flow velocity.⁹ The parameters derived from TCD such as the pulsatility index (PI) can be used to calculate ICP elevation greater than 20 mmHg: when a PI threshold of 1.3 is used, prediction of ICP values ≥20 mmHg had 100% sensitivity and 82% specificity.⁹ Since pigs have large gyrencephalic brains which would capture the complex physiological process that occurs following the initial hypoxic insult, we used a pediatric pig model of anoxic brain injury that has unique clinical relevance. The objective of this study was to have a first look at correlating contrast-enhanced ultrasound (CEUS)-derived parameters with ICP levels following cardiac arrest and ROSC.

Materials and Methods

All experimental procedures in the work were approved by the Institutional Animal Care and Use Committee. Four-week old, 10 kg pediatric swine (n = 11) were sedated using ketamine followed by 2 to 4% inhaled isoflurane. After intubation, a 3 cm cranial window was surgically made over the frontal region. A Siemens 9EC4 ultrasound transducer (Siemens Healthineers, Malvern, PA) was positioned over the cranial window with intact dura while scans were performed on an ACUSON Sequoia Ultrasound System (Siemens Healthineers, Malvern, PA). Baseline CEUS measurements were collected after an intravenous bolus of 0.2 mL of the ultrasound contrast agent Lumason® (Bracco Diagnostics Inc., Monroe Township, NJ). In the contralateral frontal region, a brain tissue ICP monitor was placed using a bolt assembly (Hemedex, Waltham, MA) and ICP was monitored throughout the experiment.

Asphyxia was induced by clamping the endotracheal tube for 7 minutes followed by electrically inducing ventricular fibrillation. This asphyxia-associated cardiac arrest model has been described in detail elsewhere.^10–12 Hemodynamic directed cardiopulmonary resuscitation CPR was performed for 10 minutes prior to the first defibrillation attempt. Starting 2 minutes into CPR, an epinephrine bolus (200 mcg) was given intravenously, as needed, with a goal of maintaining a coronary perfusion pressure (CPP) of at least 20 mmHg, as determined by the difference between simultaneously measured aortic and right atrial pressures. One minute following the initial epinephrine dose, the CPP was again used to determine whether to deliver an additional epinephrine bolus. As CPR continued and CPP fell below 20 mmHg, epinephrine and vasopressin boluses were given in a protocolized fashion (epinephrine, epinephrine, and vasopressin) as previously described.^10–12 Manual chest compressions were delivered with a target rate of 100/min with force titrated in real-time to achieve a goal systolic aortic pressure of 90 mmHg. Ventilations were delivered at a rate of 10/min with 100% oxygen. Every 2 minutes, brief interruptions in CPR to mimic in-hospital pulse checks and rhythm analyses were performed. An initial biphasic waveform defibrillation attempt (50 Joules) was provided after 10 minutes of CPR. Cardiopulmonary resuscitation and additional defibrillation attempts continued until ROSC for up to 10 additional minutes of resuscitation after the initial defibrillation.

Five Doppler samples of thalamic vessels identified by a pediatric radiologist (M.H.), and CEUS cine clips for 90 seconds during each of the 2 to 3 bolus injection were acquired 10 minutes pre-asphyxia (baseline), immediately post-ROSC, 1-hour post-ROSC, and 3-hours post-ROSC. At the end of each bolus injection, the sonographer ensured that the contrast was fully cleared out of the field of view before proceeding on to the next bolus injection. As the 5 minutes post-ROSC period is a critical time for the survival of the pigs, bolus injections were acquired providing the pigs had stable vital signs. The mechanical index was maintained between 0.10 and 0.15 during the bolus injections. Spectral waveform analysis of Doppler data was performed. Doppler ultrasound parameters (peak systolic velocity, end diastolic velocity, systolic-diastolic ratio [SDR], and resistive index [RI]) were compared with ICP and mean arterial pressure (MAP) using a linear regression to test for a linear correlation.

Image analysis was performed using MATLAB (Mathworks, Inc., Natick, MA) Image Processing Toolbox. Regions of interest (ROIs) were designated around bilateral thalami by pediatric radiologist (M.H.) from the coronal plane images of CEUS. Time-intensity curves (TICs) were generated by reporting the intensity values from ROI, and the background intensity was subtracted from all points on the TICs. The following CEUS perfusion parameters were calculated: wash-in slope, wash-out slope, early wash-out slope, time to peak (TTP), and peak enhancement (PE). PE is defined as the highest contrast intensity value, quantified by arbitrary units (a.u.). TTP was defined as the time between contrast intensity value of 20% PE and 100% PE. Wash-in slope was defined as a slope of contrast intensity value during the time of contrast intensity of 20% PE to 100% PE. Wash-out slope was defined as the slope from 100% PE to intensity returning to 20% PE. Early Wash-out slope was defined as the slope from 100% PE to intensity returning to 80% PE. Contrast images were also qualitatively assessed, with higher degree of contrast echogenicity in the ROI showing hyper-perfusion pattern and lower degree of contrast echogenicity showing hypoperfusion pattern.

The CEUS parameters were compared to ICP using a linear regression to test for a linear correlation after testing for normality and homoscedasticity. The Pearson coefficient, P-value, and standard error of slope were calculated. Moreover, an analysis of variance (ANOVA) and Tukey post hoc test were used to compare the differences between the values of the bolus perfusion metrics and hemodynamic values during each time period after testing for normalcy and equal variance.

Results

Among the 11 pigs in this study, there were 8 surviving pigs and 3 nonsurviving pigs post-arrest. The data from nonsurviving pigs were included for baseline CEUS parameter analysis, but no data were available for post-ROSC time points. Among the surviving pigs, one pig was noted to have inadvertently delayed unclamping of the endotracheal tube after arrest and another developed localized hematoma during craniotomy. Data integrity were not clearly affected by either of these findings and these animals were therefore included in the relevant analyses. Comparison of CEUS parameters and physiological parameters between surviving and nonsurviving pigs showed largely no difference, except higher TTP and ICP values in nonsurviving pigs as shown in Supplementary Table 1.

Figure 1 demonstrates a typical time course of CEUS images through a bolus at baseline, immediately post-ROSC, and 3 hours post-ROSC. Compared to the pattern of maximal contrast enhancement at baseline, the immediate post-ROSC contrast enhancement pattern shows a global hyper-perfusion pattern. At 3 hours post-ROSC, the contrast images appear qualitatively similar to the baseline patterns. The CEUS parameters used in this study are demonstrated graphically in Figure 2.

Figure 1.

Temporal evolution of brain contrast-enhanced ultrasound (CEUS)-based cerebral blood flow in a porcine model of cardiac arrest depicted with a bolus injection technique at A, baseline, B, Immediate post-return of spontaneous circulation (ROSC), and C, 3 hour post-ROSC. From left to right, CEUS images at baseline, midway through the contrast bolus, peak enhancement, midway through washout and delayed wash-out are displayed.

Figure 2.

Parameters used for contrast-enhanced ultrasound (CEUS) analysis are displayed with respect to the contrast intensity curve over time.

Although there was a significant group difference in PE, wash-in slope, and early wash-out slope, a post-hoc test showed that there was no statistically significant change from baseline among the CEUS parameters at immediate post-ROSC, 1 hour post-ROSC and 3 hours post-ROSC as shown in Table 1. However, there is a general trend of elevation immediately post-ROSC followed by decrease at 1 hour and normalization of the following parameters: PE, wash-in slope, and wash-out slope. Additionally, non-dimensional indices that indicate resistance of thalamic vasculature such as SDR and RI were significantly lowered immediately post-ROSC.

Table 1.

Summary of Statistics for the Hemodynamic, Doppler, and CEUS Parameters During Each Time Period

CEUS Parameter	Baseline	Immediate Post-ROSC	1 h Post-ROSC	3 h Post-ROSC	F (ANOVA)	P-Value (ANOVA)
PE (a.u.)	58.0±21.9	72.6±30.1	39.9±20.3	51.0±26.3	3.10	.018
TTP (s)	6.4±3.0	5.5±2.0	6.9±2.3	5.1±1.3	1.70	.178
Wash-in slope (a.u./s)	9.4±3.7	13.1±7.2	5.7±3.3	9.5±5.5	4.68	.005
Wash-out slope (a.u./s)	0.8±0.4	1.1±0.6	0.6±0.2	0.8±0.4	2.55	.650
Early wash-out slope (a.u./s)	3.6±1.3	2.5±1.2	2.4±1.4	3.6±1.6	3.55	.020
Hemodynamic parameters
ICP (mmHg)	17.0±6.5	22.4±5.0*	14.2±2.2	14.6±2.6	6.9	<.001
MAP (mmHg)	63.5±13.0	137.3±17.3**	65.5±10.7	67.0±11.9	78.8	<.001
Doppler parameters
EDV (cm/s)	7.8±4.8	30.8±22.9**	6.4±2.7	8.0±5.0	26.2	<.001
PSV (cm/s)	15.2±8.8	43.7±31.5**	11.3±4.5	16.3±10.7	17.7	<.001
SDR	2.0±0.2	1.3±0.4**	1.8±0.3	2.0±0.3	15.9	<.001
RI	0.5±0.1	0.3±0.1**	0.4±0.1	0.5±0.1	19.5	<.001

^*P< .05,

^**P < .001 compared to baseline value using post hoc test.

CEUS indicates contrast-enhanced ultrasound; EDV, end diastolic velocity; ICP, intracranial pressure; PE, peak enhancement; PSV, peak systolic velocity; RI, resistive index; ROSC, return of spontaneous circulation; SDR, systolic-diastolic ratio; TTP, time to peak.

Next, a linear regression of CEUS parameters to ICP values was performed to gain further insight into brain health after cardiac arrest as demonstrated in Table 2. At baseline, the only parameter that had significant correlation with ICP was TTP. However, at immediately post-ROSC time point, wash-out slope (Figure 3) and early wash-out slope (Supplementary Figure 1) showed significant ICP correlation. There was no significant relationship between ICP and wash-in slopes at any of the time points. At 3 hours post-ROSC, there was a significant negative correlation between ICP and PE as shown in Figure 4.

Table 2.

Linear Regression Between ICP and CEUS Parameters During Each Time Period

Period	Bolus Parameter	R 2	P value
Baseline	PE (a.u.)	0.0004	.923
	TTP (s)	0.3180	.010
	Wash-in slope (a.u./s)	0.0906	.197
	Early wash-out slope (a.u./s)	0.0161	.593
	Wash-out slope (a.u./s)	0.0388	.405
Immediate post-ROSC	PE (a.u.)	0.0646	.506
	TTP (s)	0.2964	.130
	Wash-in slope (a.u./s)	0.0033	.883
	Early wash-out slope (a.u./s)	0.6267	.010
	Wash-out slope (a.u./s)	0.6473	.009
1-h post-ROSC	PE (a.u.)	0.0250	.589
	TTP (s)	0.4410	.010
	Wash-in slope (a.u./s)	0.2518	.068
	Early wash-out slope (a.u./s)	0.2322	.081
	Wash-out slope (a.u./s)	0.0007	.927
3-h post-ROSC	PE (a.u.)	0.2768	.044
	TTP (s)	0.0743	.325
	Wash-in slope (a.u./s)	0.1875	.107
	Early wash-out slope (a.u./s)	0.2526	.056
	Wash-out slope (a.u./s)	0.0067	.772

Bolded numbers indicate statistically significant correlation values. CEUS indicates contrast-enhanced ultrasound; ICP, intracranial pressure; PE, peak enhancement; ROSC, return of spontaneous circulation; TTP, time to peak.

Figure 3.

Linear regression of wash-in slope (A–D) and wash-out slope (E–F) to intracranial pressure (ICP) is displayed over the course of 4 time points. Baseline values are shown in (A) and (E), immediate post-return of spontaneous circulation (ROSC) is shown in (B) and (F), 1 hour after ROSC is shown in (C) and (G), and 3 hours after ROSC is shown in (D) and (H). a.u. indicates artificial units.

Figure 4.

Linear regression of peak enhancement compared to intracranial pressure at baseline (A), immediate post return of spontaneous circulation (ROSC) (B), 1 hour post ROSC (C), and 3 hours post ROSC (D). a.u. indicates artificial units.

Discussion

In this initial report on CEUS parameters in the post cardiac arrest period, significant findings were made from linear regression to ICP. There was significant correlation between ICP and wash-out slope at immediate post-ROSC time point. Additionally, significant negative correlation was found between ICP and PE at 3 hours post-ROSC. Although the ICP values in the current study did not elevate much higher than the physiological range, further increases may result in prolonged CEUS parameter derangements beyond immediate post-ROSC time. Characterization of CEUS parameters in higher ICP ranges and at longer post-ROSC time points are future avenues of research, but the significant correlations of CEUS parameters to ICP in this report indicate that there is future potential in using this technology to gain insights into intracranial pathology.

The early wash-out slope also showed correlation with ICP immediately post-ROSC. As previously described, wash-in and wash-out slopes are reflective of regional blood flow velocity.¹³ The significant correlation between ICP and wash-out slope (Figure 3) and early wash-out slope (Supplementary Figure 1) immediately post-ROSC indicates that the immediate post-ROSC hyperperfusion response is associated with elevated ICP. While elevated ICP can in fact limit the cerebral blood flow in other settings, the extent of ICP elevation immediately post ROSC may not be sufficiently high enough to prevent the exaggerated reperfusion response in the context of cerebral autoregulatory dysfunction.

In contrast to the immediate post-ROSC period, other time points post cardiopulmonary resuscitation do not exhibit significant correlation between cerebral perfusion metrics and ICP as shown in Figure 3. This may be explained on the basis of intact or disrupted cerebral autoregulation. At baseline and beyond the immediate post-ROSC periods, intact cerebral blood flow autoregulation creates a stable environment in which both the inflow and outflow cerebral blood flow may not change despite changes in ICP. The mechanism behind disruption of autoregulation involving regional outflow (wash-out slope) is specifically disrupted at immediate post-ROSC period requires further investigation. One possibility of this finding is that vasoactive chemicals such as catecholamines and nitric oxide can maintain autoregulation of cerebral inflow but not outflow at this time point. It should also be noted that substantial elevations in ICP were not observed in our cardiac arrest cohort although subpopulations of long-term survival models may experience different trends in ICP. Immediately following cardiac arrest, there is likely a temporary derangement of cerebral blood flow autoregulation resulting in increasing outflow velocity with increasing ICP. In addition, there is a significant increase in MAP and decreased vascular resistance shown by RI and SDR at immediate post-ROSC (Table 1), likely to increase perfusion in the brain. This is also likely due to vasopressors administered during CPR, with increase in ICP which may reflect larger blood volume in the intracranial space. Higher ICP may also result in higher wash-out slope as the elevation in ICP with relatively unchanged systemic venous pressure may lead to higher outflow velocity.

Older studies looking at large vessels also showed that higher ICP beyond the range we report (>35–40 mmHg) reduces cerebral blood flow velocity and cerebral autoregulation is disrupted.^14–16 Although the ICP values reported in the current study are within the range of intact cerebral autoregulation, correlation in out-flow velocity with ICP is noted immediately post-ROSC. Thus, there is a temporary derangement of cerebral autoregulation which then resolves and returns to base-line at 1 and 3 hours post-ROSC. The data in this study also suggest that there are delayed changes in cerebral hemodynamics following cardiac arrest as PE shows significant inverse correlation with ICP as shown in Figure 4. At 3 hours post-ROSC, higher ICP levels correlate with reduced PE, although this parameter had no correlation to ICP in prior time points. A potential explanation of this finding may be that at this time, cytotoxic edema from anoxic injury may begin to play a significant role in cerebral blood flow. Thus, higher ICP values may start to reduce cerebral blood flow although other CEUS parameters are yet unaffected.

When CEUS parameters at various post-ROSC time points were directly compared to baseline, no statistically significant changes were found as shown in Table 1. This is likely due to the high variability in the data, although the mean values at each time point show a consistent pattern with each of the analyzed parameters. Higher trending values of PE and wash-in slope immediately post ROSC compared to baseline suggest hyperperfusion at this time point. This may be secondary to the effects of vasopressors that were administered at the time of resuscitation, which elevate the MAP to significantly higher values than baseline.

At 1 hour post-ROSC, lower PE and wash-in slope compared to baseline show the opposite pattern, suggesting a state of hypoperfusion. Similarly, wash-out slope immediately post-ROSC has higher trending values than baseline, which subsequently becomes lower than baseline at 1 hour post-ROSC. This suggests that there is a faster rate of the contrast movement into and out of the brain immediately post-ROSC, followed by a slower rate of movement at 1 hour post-ROSC. By 3 hours, the CEUS parameters return close to baseline.

There are several limitations to this study. It should be noted that the current brain CEUS imaging was performed through a burr hole to mimic a fontanelle in infants and maximize the visualization of intravascular microbubbles. Future approaches to studying cerebral blood flow using CEUS can integrate advanced post-processing algorithms to enhance acoustic signal when imaging through the bone. Such efforts will expand the clinical utility of brain CEUS to patients of all ages in need of post arrest neuromonitoring. In adults, the temporal window is oftentimes used as an acoustic window to measure macrovascular flow as in TCD examination. Brain CEUS has similarly been applied in the clinical setting to diagnose a variety of neurologic disorders in children and adults using the temporal bone as the acoustic window.¹⁷ With improvements in signal enhancement, resolution of intravascular micro-bubbles may yield important insights into not only gross tissue perfusion but also microvascular flow dynamics in the setting of cardiac arrest.

Conclusion

Our results demonstrate an important utility of brain CEUS in understanding the temporal evolution of cerebral hemodynamics and ICP post cardiac arrest, which has significant clinical implications. Measurement of ICP is often difficult given the risks involved in using invasive monitoring devices, and imaging of flow velocity as well as CEUS parameters that correlate with ICP at the same time is a potential advantage of this modality. Further understanding of the pathophysiological and functional correlates of CEUS-based cerebral blood flow parameters will undoubtedly advance the clinical utility of this noninvasive imaging tool in cardiac arrest.

Abbreviations

CEUS

contrast-enhanced ultrasound

CPP

coronary perfusion pressure

CPR

cardiopulmonary resuscitation

EDV

end diastolic velocity

ICP

intracranial pressure

MAP

mean arterial pressure

PE

peak enhancement

PSV

peak systolic velocity

RI

resistive index

ROI

region of interest

ROSC

return of spontaneous circulation

SDR

systolic-diastolic ratio

TIC

time-intensity curve

TTP

time to peak