Resuscitation with epinephrine worsens cerebral capillary no-reflow after experimental pediatric cardiac arrest: An in vivo multiphoton microscopy evaluation

Abstract

Epinephrine is the principal resuscitation therapy for pediatric cardiac arrest (CA). Clinical data suggest that although epinephrine increases the rate of resuscitation, it fails to improve neurological outcome, possibly secondary to reductions in microvascular flow. We characterized the effect of epinephrine vs. placebo administered at resuscitation from pediatric asphyxial CA on microvascular and macrovascular cortical perfusion assessed using in vivo multiphoton microscopy and laser speckle flowmetry, respectively, and on brain tissue oxygenation (PbO₂), behavioral outcomes, and neuropathology in 16–18-day-old rats. Epinephrine-treated rats had a more rapid return of spontaneous circulation and brisk immediate cortical reperfusion during 1–3 min post-CA vs. placebo. However, at the microvascular level, epinephrine-treated rats had penetrating arteriole constriction and increases in both capillary stalling (no-reflow) and cortical capillary transit time 30–60 min post-CA vs. placebo. Placebo-treated rats had increased capillary diameters post-CA. The cortex was hypoxic post-CA in both groups. Epinephrine treatment worsened reference memory performance vs. shams. Hippocampal neuron counts did not differ between groups. Resuscitation with epinephrine enhanced immediate reperfusion but produced microvascular alterations during the first hour post-resuscitation, characterized by vasoconstriction, capillary stasis, prolonged cortical transit time, and absence of compensatory cortical vasodilation. Targeted therapies mitigating the deleterious microvascular effects of epinephrine are needed.

Introduction

Brain injury occurs in greater than 50% of children who achieve return of spontaneous circulation (ROSC) after cardiac arrest (CA).^1,2 Epinephrine is the fundamental therapy that facilitates ROSC after CA, increasing aortic diastolic pressure and coronary perfusion pressure during cardiopulmonary resuscitation (CPR) through α1-mediated vasoconstriction.^3–6 Each minute of delay in epinephrine administration is associated with 4–9% lower odds of survival from CA,^7,8 and shorter time to the first administration of epinephrine is associated with favorable outcomes in pediatric patients. ⁹ A recent trial of epinephrine vs. placebo in CA showed increased ROSC and 30-day survival in patients treated with epinephrine, but no effect on survival to discharge with good neurological outcome. ¹⁰ Additionally, several experimental and clinical studies suggest that epinephrine confers no neurological outcome benefit and even potentially contributes to brain injury.^11–13 Currently, clinicians are left with the difficult choice of using epinephrine during CPR despite its potentially deleterious neurologic side effects, or performing CPR without epinephrine and consequently reducing the rate of successful resuscitation; neither choice is acceptable. There is a pressing need to characterize, target, and mitigate the deleterious effects of epinephrine in CA.

The potential detrimental effects of epinephrine on neurological outcome are incompletely characterized but have been postulated to result, at least in part, from microvascular constriction. In a classic study, Ristagno et al. used qualitative optical imaging and observed that the pial microvasculature was less visible from 3–8 min post-CA in pigs treated with epinephrine vs. placebo.^14–16 The pial cortical flow returned to normal by 10 min post-ROSC, without hypoperfusion. These authors acknowledged the limitation of the imaging method used, which restricted their observations to the surface cortical vessels, along with the absence of sustained hypoperfusion in their model, and recognized that confirmation of these results to deeper layers of the cortex and more severe CA models remains to be determined.^14–16 Thus, details of the specific microvascular compartments affected by epinephrine, quantification of the deleterious effects, and assessment of capillary flow remain unknown. A thorough characterization of the microcirculatory disturbances post-CA is needed to optimally target and mitigate epinephrine’s purported deleterious effects.

We recently defined the microvascular disturbances that produce cortical hypoperfusion in a clinically relevant model of pediatric CA that included administration of epinephrine.^17,18 This CA model produces long-lasting cortical hypoperfusion and tissue hypoxia to at least 120 min post-CA. ¹⁸ Using in vivo multiphoton microscopy of cortical microvessels to a depth of 300–400 µm, we showed that after resuscitation from CA arteriolar vasoconstriction occurs along with multifocal capillary stasis (no-reflow). ¹⁷ These microvascular disturbances are associated with impaired neuronal activity, cortical hypoxia, and impaired metabolism. ¹⁹ Epinephrine administered at resuscitation may contribute to these derangements in microvascular perfusion.

Using our established state-of-the-art approach to characterize the cerebral microcirculation with multiphoton microscopy in the pediatric CA model, our objective was to define the neurovascular effects of epinephrine. We assessed the effect of epinephrine on cerebral microcirculation, and in the same model, its effects on macrovascular cerebral blood flow (CBF), brain tissue oxygenation (PbO₂), and neurological outcomes. To avoid the indication and resuscitation biases that confound clinical studies, we used a threshold insult of 9.5 min, in which survival can be obtained with or without epinephrine. Our results suggest that epinephrine at resuscitation produces microvascular alterations during the first hour post-resuscitation, characterized by vasoconstriction, capillary stasis, prolonged cortical transit time, and absence of compensatory cortical vasodilation. These results provide a platform for the development of therapies to mitigate the negative cerebrovascular effects of epinephrine, thus allowing it to be used in a safe manner to both achieve ROSC and improve neurological outcome after pediatric CA.

Methods

Animals

Postnatal 16–18 day old male Sprague Dawley rat pups (Envigo, US), n = 64, were used in four separate experiments to assess the effect of epinephrine vs. normal saline (NS) placebo administered at resuscitation on the following parameters: cortical microvascular perfusion (n = 16), cortical perfusion measured with laser speckle flowmetry (hereto referred as macrovascular perfusion) (n = 14), cortical brain tissue oxygenation  (n = 14, same cohort as cortical macrovascular perfusion), thalamic brain tissue oxygenation (n = 12), behavioral outcomes and neuropathology (n = 22). All experiments were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Pittsburgh, and the care and handling of the rats were in accord with National Institutes of Health guidelines for ethical animal treatment. The animal data reporting of the current study has followed the ARRIVE 2.0 guidelines. ²⁰

Randomization

Rats were randomly assigned to receive either epinephrine or placebo at resuscitation from CA. A separate technician, not otherwise involved in the study, randomized the rats and prepared the medications. The syringes containing epinephrine or placebo were indistinguishable from each other. The technicians performing surgeries, CA experiments, resuscitation, microcirculatory measurements, cortical macrovascular perfusion measurements, behavioral assessments, histological analysis, and data analysis were blinded to the study group. Unblinding occurred after all data were quantified and before statistical analysis.

Anesthesia and surgery

Surgery was performed following our previously established model of asphyxial CA.^21,22 Rats were anesthetized with 3% isoflurane and 50% N₂O/balance O₂ in a Plexiglas chamber until unconscious and then tracheally intubated and mechanically ventilated. Ventilation parameters were adjusted to maintain the PaCO₂ levels within the target range of 35–45 mmHg. Isoflurane was decreased to 1.5% and central venous and arterial catheters were surgically inserted. In the experiments assessing perfusion and PbO₂, isoflurane was discontinued, and analgesia and neuromuscular blockade were achieved through the administration of fentanyl (50 µg/kg/h) and vecuronium (5 mg/kg/h), respectively, to minimize the effect of inhaled anesthetics on CBF. CA was induced by disconnecting the ventilator for 9.5 min. Resuscitation was initiated by resuming mechanical ventilation with FiO₂ = 1.0, administration of the study medication (epinephrine 0.05 mg/kg or placebo based on randomization), sodium bicarbonate (1 mEq/kg), and chest compressions until ROSC. After ROSC was achieved, anesthesia and neuromuscular blockade were restarted. FiO₂ = 1.0 was maintained for 30 min and then adjusted to FiO₂ = 0.5 for the remainder of the experiment. Rats were monitored for 60 min post-ROSC. Arterial blood gases (ABGs), mean arterial pressure (MAP), heart rate (HR), and end tidal CO₂ (ETCO₂) values were obtained at baseline (pre-CA), 10, 30, and 60 min after ROSC. For the duration of the experiment, rats were maintained normothermic (rectal temperature 37–37.5°C) through use of a heating pad and fan. Sham-operated rats underwent the same surgical procedures, but without CA, study medications, or resuscitation.

Microcirculatory measurements using multiphoton laser microscopy

We evaluated the cortical microcirculatory perfusion pre-CA and post-ROSC using in vivo multiphoton microscopy. Rats (n = 8/group) were randomized to receive epinephrine or placebo at resuscitation. For this cohort of rats, surgical tracheotomy was performed to facilitate mechanical ventilation during in vivo microscopy. Monitoring was limited to HR, MAP, and ETCO₂ with mechanical ventilation using our standard parameters to maximize the multiphoton imaging time during the first hour post-ROSC. Multiphoton microscopy data were acquired as we have previously described. ¹⁷ Briefly, rats were placed into a stereotaxic apparatus and stabilized in prone position. A circular craniotomy of 4-mm diameter was made over the left parietal cortex and the skull was removed while maintaining the dura intact. Agarose gel was applied, and a 5-mm coverslip was secured with tissue glue and dental cement to maintain visual access to the brain. To acquire in vivo microscopy data, we used a Nikon A1R MP microscope configured with the Nikon Ni-E upright motorized system, Chamelon Laser Vision, a 1.1 NA APO LWD 25X water immersion objective, and a high-speed (30 frames/s) resonant scanning mode. A Nano-Drive system was used to acquire high-speed control of z-plane selection.

Images of the cortical microvasculature were obtained pre-CA and serially from 5–60 mins post-ROSC. Fluorescein isothiocyanate-dextran (FITC, Sigma, wt 2MDa, 3% w/v), 0.1 ml, was used to label the vasculature. We used the FITC 525/50 nm (green channel) detector, an excitation wavelength of 900 nm, laser power of 1–3% and a HV range of 30–50. The microscope stage and the rats were encased in a temperature-controlled black plexiglass chamber to prevent ambient light interference. NIS Elements, Imaris, ImageJ, and MATLAB were utilized for post-acquisition image processing.

Quantification of the diameter of pial vessels, penetrating vessels and capillaries

Data were processed using the NIS element version 5.2 and ImageJ software. We imaged a field of view of 512 × 512 pixel at a resolution of 1.06 µm/pixel using 8× resonance mode and a Z stack of 350–400 µm depth. For each rat, three regions of interest (ROI) with adequate contrast resolution were identified pre-CA within the imaging window. The imaging software saves and recalls these ROIs. We acquired image stacks in the same ROI serially at 5, 30 and 60 min post-ROSC, and we analyzed these images in reference to the pre-CA image stack. The pre-CA image stack for each ROI was set as the reference stack and was overlaid with the image stacks obtained post-ROSC using the “merge” channel function. To ensure the vessels were matched in x-y-z dimensions, the “align” function was used to combine the pre- and post-ROSC images. Pial arterioles, penetrating arterioles, and capillaries (vessels with diameter <10 µm) were identified using anatomical characteristics as previously described. ¹⁷ To measure the change in the diameter of pial and penetrating vessels, the combined and aligned images were transferred to ImageJ. The validated vessel diameter plug-in described by Fischer for measuring cerebral blood vessels in rats was used to obtain vessel diameters from the maximum projection of the aligned images. ²³ To generate the mean diameter of the vessel, the width of each pial arteriole was measured at 2–3 locations, including the widest and narrowest diameters of the segment, and the mean value was recorded. For penetrating arterioles, the cross-sectional surface was measured in two perpendicular axes to obtain the mean diameter. Capillary diameters were quantified from flow videos using scripts written in Python to conduct the same analysis in more vessel segments. To identify the vessel boundaries, we used the intensity profile of the bright vessel and dark background. The vessel boundary was recorded as the point where the intensity profile reached half of the maximum value from both sides of the profile (a value of 0.5 on the normalized profile). The vessel diameters at 5, 30 and 60 min post-ROSC were reported as the percent change from pre-CA values.

Quantification of capillary red blood cell (RBC) flow as a direct measure of capillary stalling

To assess capillary RBC flow vs. stalling we recorded time-series images at a speed of 30 frames/s over 15 s and quantified the passage of RBCs through single-branched capillaries with diameter <10 µm. Three capillary branches from the ROIs were selected for each rat. Time-series images of these capillaries were acquired pre-CA and 5, 30 and 60 min post-ROSC. The capillary lumens displayed green FITC fluorescence, and the RBCs were identified as 3–6 µm black (negative) signals traversing the capillary lumens. Intravascular RBC stalling was defined as the absence of RBC passage for 10 sec within the identified capillaries that had normal RBC passage pre-CA (Supplemental Video 1). We calculated % RBC stalling = number of capillaries with RBC stalling x100/number of capillaries assessed.

Quantification of mean transit time (MTT) as an indirect measure of capillary stalling

Assessment of the MTT of a vascular label describes the overall perfusion of larger cortical areas. Prolonged MTT and broader venous intensity signal reflect reduced perfusion and more heterogenous blood transit, respectively. ¹⁷ We calculated MTT as previously described. ¹⁷ Briefly, ROIs were placed on a pial artery and an adjacent pial vein within the imaging window and time-series videos were recorded at the speed of 30 frames/s in resonance mode while an intravenous bolus of 0.1 ml of FITC dextran was administered (3% w/v). The arterial and venous intensity peaks were computed. We measured MTT (peak and spread) pre-CA and at 30 and 60 min post-ROSC and analyzed the data using the NIS Element software. ¹⁷ Measured values at 30 and 60 mins were reported as percent change from pre-CA values.

Cortical cerebral perfusion measurements using laser speckle flowmetry

To quantify macrovascular cortical perfusion, we acquired laser speckle perfusion images of the sensorimotor cortex in n = 7 rats/group, randomized to receive epinephrine or placebo at resuscitation. We assessed perfusion continuously from pre-CA and until 60 min post-ROSC. Images were obtained through the intact skull after deflecting the scalp as previously described. ²⁴ Perfusion values were generated using the PIMSoft Perimed software and were reported as percent change from pre-CA values.

Brain tissue oxygenation assessment

We quantified cortical and thalamic PbO₂ in rats randomized to receive epinephrine or placebo at resuscitation from CA. To maximize the use of rats for our experiments, cortical PbO₂ was measured in the same cohort of rats that had perfusion assessment with laser speckle flowmetry (n = 7/group), while thalamic PbO₂ was assessed in a separate cohort (n = 6/group). PbO₂ was measured using a Clark type tissue microelectrode (Ox-50, Unisense, Denmark) as previously described. ¹⁹ Briefly, a 2-mm burr hole was drilled in the left side of the skull, 2-mm lateral and 3.3-mm posterior to bregma. The microelectrode was inserted to a depth of 1-mm for cortical PbO₂ measurements and a depth of 6-mm for thalamic PbO₂ measurements. PbO₂ levels were recorded continuously pre-CA and until 60 min post-ROSC.

Behavioral analysis

Rats (n = 22) were randomized to one of the following groups: CA treated with epinephrine (CA-epinephrine, n = 8), CA treated with placebo (CA-placebo, n = 8), and sham (n = 6). Motor and cognitive function were assessed in these groups of rats.

Motor testing

The beam-balance (BB) task assesses vestibulomotor function and consists of placing the rats on a suspended, narrow (1.3 cm wide) elevated wooden beam and recording the time (maximum of 60 s) before they lose their balance and fall. The test was performed on post-surgery days 1–5 and three trials were provided (30 s inter-trial interval). The average of the three trials was used in the statistical analyses. The inclined plane test assesses grip strength ²¹ and consists of placing the rat on a plane with evenly spaced paw grips and variable inclination angles. The test was performed on post-surgery days 1–5. Each day testing begins at 45°, and if the rat manages to hold on for at least 10 s, the angle is increased by 5° until the rat either fails to hold or reaches the maximum angle of 80°. The average of the three trials was used in the statistical analyses.

Cognitive function testing

The Morris water maze (MWM) is a well-validated test to assess spatial learning and memory retention in normal and injured rats.^21,22,25,26 Briefly, rats are placed into a pool (180-cm diameter, 60-cm depth, 20–22°C) that resides in a room with permanent visual cues on each of the four walls containing an escape platform that is in the southwest quadrant of the pool and kept constant for all trials. Each of the visual cues is unique and enables the rats to orient in space and navigate towards the platform. Four daily trials with randomized entry locations (north, east, south, and west) and a maximum exploration time of 120 s/trial was provided. The mean of the four daily trials was used for the analysis. The rats were acclimated to the goal of the test on day 8–10 using a visible platform. The platform was raised 2-cm above the water surface and was marked by white tape to enable the rats to see the platform. This was followed by four days of hidden platform (day 11–14), where the platform was submerged below the water. ²⁵ On day-15 post-surgery, the rats received a probe trial without the escape platform to evaluate memory retention. The concept of the probe trial is that rats that have learned the location of the platform will spend more time in the quadrant of the pool where the platform was previously located (i.e., target quadrant) suggesting adequate memory retention. Thus, after removing the escape platform the rats swam freely for 30 s, and the percent time spent in the target quadrant was recorded. All data were recorded with AnyMaze software version 5.10 (Any-maze, Wood Dale, IL).

Histologic assessment

Following behavioral testing, rats (postnatal day 35) were anesthetized with 3% isoflurane and 1:1 O₂ and N₂O, underwent transcardiac perfusion with 50 mL heparinized saline and the brain was extracted. The left hemisphere was fixed in 10% phosphate buffer formalin for 24 h and then cryoprotected in 30% sucrose until submerged. The tissue was then flash frozen and stored at −80°C until processing. Coronal sections of 30 µm were cut on the cryostat (Leica, Buffalo Grove, IL, USA) for histology. The sections were stained with cresyl violet, and viable CA₁ dorsal hippocampal neurons were quantified by a technician blinded to the group assignment using live counting with Nikon 90i and NIS Elements AR 3.22.15 software (Nikon Incorporated, Melville, NY, USA). The length of the CA₁ region was also quantified, and CA₁ neuronal counts were computed by dividing the morphologically intact neurons by the length of the CA₁ region. For histological quantification, n = 19 of the collected tissue was utilized with CA-Epinephrine (n = 7), CA-Placebo (n = 7), and sham (n = 5).

Statistical analysis

Descriptive statistics were tabulated to summarize physiological variables with mean and standard deviation (SD) presented for continuous variables. The duration of CPR was analyzed with an unpaired t-test. Cerebral perfusion, brain tissue oxygenation changes, behavioral scores, and physiological variables were analyzed for time effect within each experiment between groups using repeated-measures analysis of variance (RM ANOVA), followed by Bonferroni post-hoc comparison. Changes in vessel diameter, capillary blood flow, and capillary MTT from pre-CA were tested between groups with mixed model using random coefficient with 100 bootstrapping. Histological data were analyzed using ANOVA with Bonferroni post-hoc comparison. We used Shapiro Wilk (SW) to assess for normal distribution. For data analyzed using repeated measure analysis, we used bootstrapping to perform the hypothesis testing and calculated the final p value, a robust methodology to overcome departure from normal distribution. For all statistical analyses, p < 0.05 was considered statistically significant. Data were analyzed with the statistical software StataSE 16.1 (StataCorp, College Station, TX, USA) and GraphPad Prism 8 (GraphPad Software, Inc., La Jolla, CA, USA). All figures were generated using GraphPad Prism 8 (GraphPad Software, Inc., La Jolla, CA, USA).

Results

Physiology

Systemic physiological and blood gas data, combined from macrocirculatory perfusion, tissue oxygen, and behavioral experiments, are presented in Table 1. Temperature, HR, pH, PaO₂, and PaCO₂ were similar for epinephrine and placebo groups pre-CA and 10–60 min post-ROSC. MAP was increased at 10 min post-ROSC in epinephrine vs. placebo (p < 0.05). Physiological data for each of the four experiments are presented in the Supplemental Table 1.

Table 1.

Physiological parameters (temperature, HR, MAP, pH, PaCO₂, PaO₂) in rats that received epinephrine or placebo at resuscitation.

	Temperature (°C)		HR (bpm)		MAP (mmHg)		pH		PaCO2 (mmHg)		PaO2 (mmHg)
Time-points	EPI	Placebo	EPI	Placebo	EPI	Placebo	EPI	Placebo	EPI	Placebo	EPI	Placebo
Pre-CA	37.3 ± 0.2	37.2 ± 0.2	359 ± 40	368 ± 36	59 ± 10	57 ± 7	7.35 ± 0.06	7.35 ± 0.05	39 ± 5	38 ± 3	203 ± 25	219 ± 10
10 min post-ROSC	36.9 ± 0.6	36.7 ± 0.9	365 ± 40	376 ± 48	77 ± 12*	66 ± 11	7.29 ± 0.10	7.32 ± 0.09	36 ± 5	35 ± 4	358 ± 74	404 ± 41
30 min post-ROSC	37.3 ± 0.3	37.3 ± 0.3	376 ± 45	383 ± 34	55 ± 10	48 ± 8	7.40 ± 0.06	7.40 ± 0.07	42 ± 6	39 ± 5	372 ± 60	401 ± 29
60 min post-ROSC	37.2 ± 0.3	37.1 ± 0.3	349 ± 35	354 ± 31	53 ± 9	48 ± 8	7.42 ± 0.04	7.42 ± 0.08	41 ± 4	39 ± 3	307 ± 115	337 ± 103

Physiological parameters from all experimental groups except for rats undergoing assessment of microcirculation are included. n = 21 rats/group. Values are represented as mean ± SD. *p < 0.05 epinephrine (EPI) vs. Placebo.

There was no difference between the weights of the two CA groups at baseline and during the 15 days post-CA (data not shown). The duration from last cardiac pulse to the onset of CPR was similar in the two groups (320 ± 94 vs. 329 ± 57 s, epinephrine vs. placebo, respectively, p = 0.7).

Epinephrine-treated rats required shorter CPR duration to achieve ROSC vs. placebo

ROSC was achieved in all rats. The duration of CPR required to achieve ROSC was shorter in epinephrine-treated rats: 32 ± 12 vs. 63 ± 22 s for epinephrine vs. placebo, respectively, p < 0.05. The CPR duration was 41 ± 15 vs. 67 ± 32 s for the microcirculation group (epinephrine vs. placebo, p = 0.06), 34 ± 12 vs. 63 ± 26 s for the macrovascular perfusion group (epinephrine vs. placebo, p < 0.05), 27 ± 2 vs. 67 ± 20 s for the PbO₂ group (epinephrine vs. placebo, p < 0.05), and 24 ± 3 vs. 58 ± 9 s for the behavioral group (epinephrine vs. placebo, p < 0.05).

Penetrating arteriolar constriction occurred in epinephrine-treated rats post-ROSC. Capillary dilation occurred only in the placebo group

The diameters of pial, penetrating arterioles, and capillaries (vessels with diameter <10 µm) were assessed using in vivo multiphoton microscopy (Figure 1(a)). Pial arteriolar constriction was observed post-ROSC in both groups and was similar for epinephrine and placebo (p = 0.64, Figure 1(b) and (e)). Penetrating arteriolar constriction post-ROSC occurred only in the epinephrine-treated rats and not in placebo (p < 0.05 epinephrine vs. placebo, Figure 1(c) and (e)). Capillary diameter remained relatively stable post-ROSC in epinephrine-treated rats, whereas the placebo group had progressive increase in capillary diameter post-ROSC (p < 0.05 placebo vs. epinephrine, Figure 1(d) and (e)). These results indicate that administration of epinephrine at resuscitation is associated with constriction of penetrating arterioles and absence of capillary dilation post-ROSC.

Figure 1.

Diameters of pial, penetrating arterioles, and capillaries in epinephrine- and placebo-treated rats at baseline (pre-CA) and post-ROSC. (a) Representative image of the cortical microcirculation acquired with multiphoton microscopy. (b) Pial arterial vasoconstriction was similar in the two groups. (c) Penetrating arteriolar vasoconstriction occurred only in epinephrine-treated rats. (d) Capillary diameters were increased post-ROSC at 60 min in the placebo group and (e) Representative images of pial, penetrating arterioles, and capillaries in the epinephrine and placebo groups pre-CA (green pseudo color) and post-ROSC (red pseudo color). The overlay images illustrate the degree of vasoconstriction (green colored, white arrow pointing towards the vessel) or vasodilation (red colored, white arrows pointing away from the vessel). n = 8 rats/group, *p < 0.05 epinephrine vs. placebo. Analyzed by mixed model using random coefficient with 100 bootstrapping.

Capillary stalling was more pronounced in epinephrine vs. placebo post-ROSC. Capillary MTT showed a greater increase in epinephrine vs. placebo

Capillary stalling was assessed using two complementary methods: RBC flow in individual capillaries and MTT of plasma through the cortical capillary bed over the imaged area. Stalling of RBC flow in capillaries occurred post-ROSC in both groups of rats. Epinephrine-treated rats had stalling in 29 ± 15%, 26 ± 18%, and 23 ± 18% of capillaries at 5, 30, and 60 min post-ROSC. Placebo-treated rats had stalling in 27 ± 17%, 15 ± 13%, and 11 ± 9% of capillaries at 5, 30, and 60 min post-ROSC (p < 0.05 epinephrine vs. placebo, Figure 2(a) and (b)).

Figure 2.

Cortical capillary flow assessment. (a) Representative image of normal capillary flow (pre-CA) and capillary stasis (post-ROSC). (b) Percent of capillaries with stalling was increased post-ROSC in epinephrine vs. placebo. 48 capillary branches were assessed for each group. (c) Mean transit time (MTT) Intensity-time curves and g fit curves from one representative rat treated with epinephrine or placebo. MTT increased post-ROSC in the epinephrine-treated rat vs. placebo, indicated by the black arrows and (d) Cortical MTT of plasma, % baseline. MTT was increased in epinephrine vs. placebo. n = 8 rats/group, *p < 0.05 epinephrine vs. placebo. Analyzed by mixed model using random coefficient with 100 bootstrapping.

In epinephrine-treated rats, capillary transit time of plasma increased post-CA to 177 ± 36% and 146 ± 48% of baseline at 30 and 60 min, respectively. In placebo-treated rats capillary MTT increased to 113 ± 31% and 130 ± 37% of baseline at 30 and 60 min, respectively (p < 0.05 epinephrine vs. placebo, Figure 2(c) and (d)).

Epinephrine-treated rats had greater increase in cortical reperfusion immediately after ROSC vs. placebo

Laser speckle flowmetry was used to measure global cortical perfusion continuously, including during CA and CPR. Immediately after CPR, epinephrine-treated rats had greater increase in cortical perfusion vs. placebo at one minute (61 ± 13 vs. 34 ± 8%, p < 0.05), two minutes (68 ± 13 vs. 49 ± 9%, p < 0.05) and three minutes post-ROSC (87 ± 8 vs. 69 ± 14%, p < 0.05). At 4 and 5 min post-ROSC cortical perfusion did not differ between the two groups (Figure 3(a) and (b)). MAP was higher at 1 and 2 min post-ROSC in epinephrine compared to placebo (Figure 3(c), p < 0.05) and not different at 3–5 min. Supplemental Video 2 illustrates time lapse videos of cortical perfusion at baseline, during CA, and during reperfusion in one representative rat from each group. From 5–60 min post-ROSC, cortical perfusion in both epinephrine and placebo was lower than baseline and did not differ between groups (Figure 3(d)).

Figure 3.

Cortical perfusion and mean arterial pressure (MAP) in epinephrine- and placebo-treated rats pre-CA, immediately post-ROSC (a–c), and from 5–60 min post-ROSC (d). (a) Cortical perfusion was higher in the epinephrine group from 1–3 min post-ROSC. (b) Images of cortical perfusion acquired by laser speckle flowmetry from representative rats treated with epinephrine and placebo pre-CA and 1–5 min after ROSC. Higher perfusion is seen in the epinephrine-treated rat from 1–3 min post-ROSC. (c). MAP was higher in the epinephrine group at 1- and 2-min post-ROSC and (d) Cortical perfusion at 5-, 30-, and 60-min post-ROSC were similar in the epinephrine and placebo groups. n = 7 rats/group, *p < 0.05 epinephrine vs. placebo. Analyzed by RM ANOVA, followed by Bonferroni post-hoc comparison.

Brain tissue oxygenation was similar in epinephrine- and placebo-treated rats

Cortical and thalamic PbO₂ were similar in the two groups of rats (p > 0.05). Figure 4(a) and (b) present the cortical and thalamic PbO₂ values from pre-CA to 5 min post-ROSC, and Figure 4(c) and (d) present the cortical and thalamic PbO₂ values from 5–60 min post-ROSC. Cortical PbO₂ decreased from pre-CA values of 77 ± 14 and 68 ± 10 mmHg in epinephrine and placebo groups, respectively, to 6 ± 7 and 10 ± 18 mmHg at 60 min post-ROSC (p > 0.05, epinephrine vs. placebo). Thalamic PbO₂ markedly increased from pre-CA values of 73 ± 48 mmHg in the epinephrine group and 63 ± 49 mmHg in placebo to 274 ± 166 and 273 ± 96 mmHg at 5 min post-ROSC, and then decreased to 27 ± 28 and 10 ± 3 mmHg at 60 min post-ROSC (epinephrine vs. placebo, respectively, p > 0.05).

Figure 4.

Cortical and thalamic brain tissue oxygenation (PbO₂) in epinephrine- and placebo-treated rats pre-CA, immediately post-ROSC (a, b), and from 5–60 min post-ROSC (c, d). Cortical (a) and thalamic (b) PbO₂ were similar in epinephrine and placebo groups from 1–5 min post-ROSC. Cortical PbO₂ (c), and thalamic PbO₂ (d), at 5-, 30-, and 60-min post-ROSC were similar in the epinephrine and placebo groups. n = 6–7 rats/group. Analyzed by RM ANOVA, followed by Bonferroni post-hoc comparison.

Motor performance was similar in rats treated with epinephrine and placebo post-ROSC

Both CA-Epinephrine and CA-Placebo groups performed worse on beam-balance vs. sham controls (p < 0.05). No difference in beam-balance performance was observed between CA-Epinephrine and CA-Placebo groups (Figure 5(a)). Similarly, on the inclined plane, both CA groups performed worse than shams on day 1 post-CA (p < 0.05). On days 2-5 CA groups performed similar to shams (Figure 5(b)). No difference in inclined plane performance was observed between CA-Epinephrine and CA-Placebo groups.

Figure 5.

Functional outcome assessment. (a) On beam-balance, cardiac arrest (CA)-Epinephrine and CA-Placebo performed worse than shams on days 1–3. (b) On the inclined plane, CA-Epinephrine and CA-Placebo performed worse vs. shams on day 1. Memory acquisition on the visible (c) and hidden (d) platform tests showed similar improvements in learning trials for CA-Epinephrine CA-Placebo vs. sham and (e) Memory retention assessed using the probe trial showed impaired memory retention in rats treated with epinephrine vs. shams, whereas placebo-treated rats performed similarly to shams. *p < 0.05 CA-Epinephrine vs. shams. ^#p < 0.05 CA-Placebo vs. shams. CA-Epinephrine and CA-Placebo n = 8 rats/group, sham group, n = 6 rats/group. Analyzed by RM ANOVA, followed by Bonferroni post-hoc comparison.

Spatial memory acquisition in the MWM test was similar for epinephrine and placebo. Epinephrine-treated rats performed worse than shams in the memory retention task

There was no difference between CA groups and shams on the visible platform task, indicating that all the rats had adequate vision and were similarly acclimated to the MWM test (Figure 5(c)). During the hidden platform testing, there was no overall difference between the groups (p = 0.1) (Figure 5(d)). Analysis of the probe trial data showed worse memory retention for CA-Epinephrine group, as demonstrated by a shorter percentage of the 30 s allotted time spent in the target quadrant for the CA-Epinephrine group vs. sham (p < 0.05), relative to the CA-Placebo group which did not differ from shams (p > 0.05, Figure 5(e)). There was no difference between the CA groups. Swim speeds were not different between groups.

Hippocampal neurons were decreased in both CA groups vs. shams, with no difference between CA epinephrine and CA placebo post-ROSC

Neuronal counts in the CA₁ region of the dorsal hippocampus were lower after CA vs. shams. CA-Epinephrine rats had 255 ± 55 neurons/mm, CA-Placebo rats had 225 ± 56 neurons/mm, whereas shams had 371 ± 79 neurons/mm, p < 0.05 for each CA group vs. sham, and p > 0.05 for CA-Epinephrine vs. CA-Placebo.

Discussion

We report the first comprehensive assessment of the microcirculatory effects associated with the administration of epinephrine at resuscitation from CA. In our pediatric CA model, epinephrine exacerbates the microcirculatory dysfunction post-CA, producing additional penetrating arterial vasoconstriction, absence of compensatory capillary dilation, and worsening capillary stasis (no-reflow) in the first 30–60 min after ROSC vs. placebo. Our study also suggests that the vasoconstricting effects of epinephrine on the microcirculation may not be apparent in our assessments of either the macrocirculation or brain tissue oxygenation. Administration of epinephrine did not improve the neurological outcome in the experimental pediatric asphyxial CA model. As withholding epinephrine is not an acceptable strategy in CA, our study provides a platform for the development of strategies to mitigate the unwanted microvascular effects of epinephrine to optimize its use during resuscitation.

Epinephrine is the principal medication administered during CPR to improve ROSC. Its benefit during CPR to attain ROSC is indisputable,^27–30 and delaying its administration decreases the rate of ROSC.^7,8 Our study highlights two aspects of the blood flow promoting effect of epinephrine in the heart and the brain: shortened duration of CPR to attain ROSC, along with immediate and brisk reperfusion post-ROSC, both effects shortening the duration of cerebral hypoperfusion post-CA. The increase in cortical perfusion during the first three minutes after CPR observed using laser speckle flowmetry in rats treated with epinephrine agrees with many studies showing increased cortical perfusion during CPR and immediately after ROSC in various species of adult and pediatric ages, after immediate or delayed administration of epinephrine.^{3–6,31–33} These blood flow promoting effects are also supported by the results of a recent metanalysis of studies in pediatric CA suggesting that early administration of epinephrine at 3, 5, or 10 min post-CA is associated with improved ROSC, neurological outcome, and survival to hospital discharge vs. late epinephrine administration at >15 min post-CA. ⁹

It is unclear at this time if the rapid reperfusion produced by epinephrine is a) beneficial, rapidly providing flow to the cerebral ischemic tissue, b) neutral, simply reflecting increased MAP in the context of impaired autoregulation, c) detrimental, promoting endothelial damage, arteriolar vasoconstriction, and/or oxidative injury, or d) producing a combined effect, initially beneficial and then detrimental to the microcirculation. Given our data, it is tempting to speculate that the latter combined effect of initial benefit/later detriment is at play. One possibility is that the rapid reperfusion and the supra-normal wall tension induced by epinephrine could lead to a resetting of the basal set point for the arteriole’s myogenic tone and account for the constriction of penetrating arterioles and consequent capillary stasis. Epinephrine-induced oxidative injury at the level of the endothelium may also be responsible for the vasoconstriction and capillary stasis observed using in vivo microscopy.^34,35

In our study, the initial resuscitation dose of epinephrine produced protracted vascular effects seen at 30-60 min post-CA. The cerebral microvascular derangements have been postulated to occur secondary to the vasoconstrictor α₁ effect of epinephrine, however the α₁ effect would not be expected to last for 30-60 min after the administration of a single post-resuscitation dose. The combination of penetrating arteriolar constriction and absence of capillary dilation, together with capillary stasis, could possibly worsen the secondary ischemic insult to the cortex.

Several mechanisms could potentially underlie the protracted microcirculatory effects of epinephrine in CA, acting alone or in combination. Epinephrine could increase arteriolar wall tension and produce subsequent vasoconstriction and capillary stasis independent of the α₁ effect. Epinephrine may worsen the oxidative injury in the endothelium and vascular contractile cells, inducing vasoconstriction and contributing to the capillary stasis.^34,35 During CA there may be inhibition of endothelial monoamine oxidase^36,37 which would allow epinephrine to have access to the endothelium, vascular smooth muscle cells, and pericytes even in the absence of blood brain barrier disruption. Epinephrine could also independently inhibit nitric oxide. ³⁸ The prothrombotic action of epinephrine may also worsen the capillary stasis.^39–41 Moreover, epinephrine may induce inflammatory changes ⁴² or apoptosis via the ß₁-cAMP interaction. ⁴³ We and others showed that antioxidant strategies administered concomitantly with epinephrine (polynitroxyl albumin, superoxide dismutase, deferoxamine) decreased the early regional hyperemia seen in subcortical areas after CA, with beneficial effects on neurological outcome.^25,44,45 Targeting oxidative stress or administering microvascular dilators such as nitrite or adenosine, inhibition of CYP vasoconstrictors, or strategies to improve capillary stasis, among others, merit further evaluation. In contrast to a study in focal cerebral ischemia that suggested pericapillary contractile cells die in rigor and constrict the capillary, ⁴⁶ our results show progressively increased capillary diameters after resuscitation from CA in the placebo group, which suggest that targeting microvascular constriction after resuscitation is achievable. The mechanisms responsible for the microcirculatory effects of epinephrine deserve further study.

Importantly, neither laser speckle flowmetry nor PbO₂ sensors were able to detect the delayed difference between epinephrine and placebo seen in the microcirculation post-CA. This may reflect the difference in the sensitivity of these methods. Laser speckle flowmetry assesses a combination of macrovascular and microvascular flow. It is possible that the fine signal of capillary flow disruption by epinephrine is lost in the more intense neutral signal of the larger pial arterioles. Moreover, due to the plane of acquisition, laser speckle is relatively insensitive to the signal of the penetrating arterioles. Likewise, PbO₂ measurement did not detect differences between the groups, although it correctly detected tissue hypoxia. ¹⁹ The absence of a difference in PbO₂ may have been either correctly identified or may reflect the method used to obtain data. In our model, cortical hypoxia post-CA reaches critical PbO₂ levels of <20 mmHg, which might be too low to detect a small additional difference in oxygenation produced by further changes in microvascular perfusion. Also, PbO₂ is measured in a small area of the cortex (50 µm) around the glass electrode. As only 25% of cortical vessels have no-reflow in our model, the PbO₂ electrode might or might not have been in proximity to an area of no-reflow. Moreover, the PbO₂ values vary depending on the distance between the electrode and an arteriole. Thus, multiple factors could dampen possible differences in PbO₂.

Motor outcome was similar in rats randomized to epinephrine vs. placebo. Rats resuscitated with epinephrine, however, showed impairment vs. sham on the probe trial—suggesting impaired cognitive function. Based on studies in this threshold insult, where behavioral and histological findings are modestly impaired, we cannot state with confidence that epinephrine worsened outcome after CA, although we can say that the outcome was not improved in epinephrine vs. placebo. The histological hippocampal neuronal counts were similar in the two groups. This threshold insult is insufficient to produce robust neuronal death in the cortex to compare histological effects in that region. The long-term functional detriment of epinephrine may or may not be directly related to the vascular microcirculatory derangements. Other studies suggest worse neurological outcome or no improvement after epinephrine, consistent with our study.^47,48

Our study has several limitations. Creating an animal model with high translational value is challenging. Our clinically relevant model simulates pediatric CA; it uses developing 16–18-day old rats, uses asphyxia as the precipitating factor for CA, and employs resuscitation with ventilation and chest compressions. This model produces selective neuronal death in the cortex, hippocampus, and cerebellum, along with long-term behavioral deficits. We recognize that our pre-clinical model has limitations. Relevant to this study, the CA duration is often less than that observed in children, especially those resuscitated from out of hospital CA. Resuscitation is also obtained after a shorter duration of CPR than is commonly seen in humans. The epinephrine dose administered, 0.05 mg/kg, is consistent with the dose used in resuscitation from CA in rats but is higher than the 0.01 mg/kg recommended in clinical pediatric care. We administered bicarbonate at resuscitation, also consistent with rodent models of CA, but not routine clinical care. This was administered in both the epinephrine and placebo groups. We used only males. As the oxidative CBF effects were more pronounced in male rats in our previous studies, assessment of microcirculatory effects of epinephrine in female rats is warranted. ²⁵ We limited our study to pediatric asphyxial CA. Assessment of microcirculatory effect of epinephrine in adult models of ventricular fibrillation and asphyxial CA is warranted, since important differences in reperfusion are seen in these models. ⁴⁹ Paradigms of early vs. late administration of epinephrine and various strategies of epinephrine administration as well as the optimal initial and subsequent epinephrine doses should be assessed. We assessed the no-reflow phenomenon in the parietal cortex, but assessment of microcirculatory changes in other cortical regions (frontal, occipital), as well as in subcortical areas such as thalamus, where the PbO₂ and CBF increase markedly early post-CA is warranted. We assessed the neuronal counts in the hippocampus—which is the most selectively vulnerable region and is routinely used to demonstrate histological correlates of cognitive outcome in pre-clinical studies. Given our findings in the cortical microcirculation, neuropathological assessments in the cerebral cortex are warranted, particularly in studies using longer arrest durations—where cortical damage is robust.

At the insult duration selected, our experimental CA model allowed us to assess of the effect of epinephrine vs. placebo in the absence of two critical biases that confound clinical studies and favor placebo: resuscitation and indication bias. The resuscitation bias occurs in randomized blinded studies and is present if predominantly subjects with shorter CA achieve ROSC in the placebo group, while a greater number of patients with longer CA achieve ROSC in epinephrine-treated group. Indication bias, which also favors placebo treatment, is present in non-randomized studies, and occurs when more severe insults are more likely to be treated with epinephrine. These biases prevent the ability to discern if the worse neurological outcome is secondary to epinephrine or to increased insult duration. ⁵⁰ To ensure that our results were not affected by the resuscitation bias, we identified a threshold insult duration of 9.5 min that allowed ROSC and long-term survival to be achieved in all rats whether epinephrine was given or not at CPR. To assure our results were not affected by the indication bias we also randomized the intervention and blinded the researchers who performed the surgery, resuscitation, treatment, and outcome analysis.

In summary, epinephrine offers significant short term hemodynamic benefits in CA, however these benefits may come at the expense of increased microcirculatory disturbances and potential long-term adverse outcome. As withholding epinephrine at resuscitation would lead to a higher proportion of patients who do not achieve ROSC, strategies to mitigate the effects of epinephrine on the neurovascular unit are needed to be able to use epinephrine in a safe manner and obtain ROSC without worsening the microvascular hypoperfusion. These strategies are important especially in the context of the current improvements in the critical care, advances in improved outcomes with the institution of targeted temperature management, and increased incidence of CA in populations where the brain is potentially neuroplastic, such as in young adults who suffer drug overdoses or in children, where favorable outcomes have been described after prolonged resuscitation. We identified the microvascular derangements produced by epinephrine which can be targeted with the goal of improving outcome. Our study provides a powerful platform to assess innovative strategies that may mitigate the deleterious microvascular effects of epinephrine.