Epinephrine and Cerebral Blood Flow in Cardiac Arrest
“I am a necessary part of an important search to which there is no end.”
With PARAMEDIC-2, the role of epinephrine in cardiac arrest, and especially its effects on cerebral perfusion and neurological outcome, have been appropriately questioned. Recall that Perkins and colleagues studied epinephrine in out-of-hospital cardiac arrest in a very well-performed randomized controlled trial of over 8000 patients. Epinephrine significantly improved 30-day survival with a number needed to treat of 112. Importantly, however, about 1/3rd of the survivors in the epinephrine group did so with modified Rankin scores of 4 or 5 – indicating severe neurological disability; only 17% in the placebo group survived with comparably poor neurological outcome. Accordingly, there is fear that epinephrine improves central hemodynamics but vitiates neurological perfusion. Amplifying this concern is data published by Ristagno and colleagues that revealed diminished cortical microvascular perfusion with epinephrine in cardiac arrest.
With this background, a porcine investigation by Mavroudis et al. as a pediatric model of cardiac arrest was recently published. The authors employed multiple invasive and non-invasive measures of cerebral flow during cardiac arrest, specifically measuring the effect of epinephrine boluses on these parameters.
What They Did
Twenty, one-month old female swine were used in this model. Cardiopulmonary resuscitation [CPR] at standardized depth at 100 compressions per minute was started immediately after induction of cardiac arrest by ventricular fibrillation. Brief, 4 second interruptions in CPR every 2 minutes mimicked pulse-checks/rhythm analysis.
20 mcg/kg of epinephrine was given 2 minutes into CPR and 4 minutes thereafter, consistent with Pediatric Advanced Life Support Guidelines. After 10 minutes of CPR, first defibrillation was performed. CPR was continued until sustained return of spontaneous circulation [ROSC] of more than 20 minutes was achieved or after an additional 10 minutes of resuscitation post initial defibrillation attempt – after which resuscitation was stopped [i.e., 20 minutes maximum]. If ROSC was achieved, the animals received protocolized ICU level care for 4 hours and then euthanized.
Both cerebral blood flow and oxygenation was measured by two, separate invasive methods and a single, non-invasive method. The two invasive techniques were direct assessment of white matter tissue oxygenation [PtO2] via a probe placed though a burr hole. The second invasive approach was trans-dural laser Doppler assessment of blood flow through a separate burr hole. Further, the non-invasive technique employed a newer embodiment of near infrared spectroscopy [NIRS] called ‘frequency domain’ NIRS. This technology measures hemoglobin, oxy-hemoglobin and total hemoglobin concentrations, brain tissue oxygenation and cerebral blood flow through the cranium.
What They Found
With the first dose of epinephrine, relative cerebral blood flow [rCBF] and tissue oxygenation significantly increased by about 10% whether measured invasively or non-invasively. By the third dose, there was no significant difference in rCBF or oxygenation. Interestingly, invasive tissue oxygenation increased with the first epinephrine dose, but not thereafter, whereas non-invasive tissue oxygenation increased with first and second dose but not third.
With systemic hemodynamics, systolic and coronary perfusion increased with the first two doses of epinephrine while diastolic pressure effects were noted through four doses of epinephrine.
Measuring blood flow is a challenging endeavour; measuring its response within a particular clinical or experimental context is muddied by methodology, model, paradigm and the specific tissue through which the blood volume of interest moves. All of these variables are dynamic and interdependent; it is unsurprising that disparate results are reported in the literature. Importantly, this doesn’t mean that one result is incorrect. As the authors note, many other investigators have confirmed that cerebral blood flow rises in response to epinephrine in cardiac arrest models. How is this reconciled with the findings of Ristagno and colleagues?
Ristagno et al. used orthogonal polarization spectral imaging-based measurements in capillaries less than 20 micrometers in diameter. With this, epinephrine diminished the microvascular blood flow index – a score where 0 represents “no flow;” 1, “markedly reduced flow;” 2, “reduced flow,” and 3; “normal flow.” As noted, this methodology only assesses highly-localized, very superficial vessels on the surface of the cerebral cortex.
Additionally, Ristagno and colleagues waited 3 minutes after the onset of ventricular fibrillation before they started their resuscitation protocol, while Mavroudis et al. started immediately. Thus, there is a 2-minute delay in epinephrine provision in the former versus latter investigations. As the grey matter of the cortex has a greater blood flow, a lower ischemic threshold and is more susceptible to glutamate-mediated excitotoxicity, its superficial-most layer could be more disposed to prolonged ischemia and perhaps untoward microvascular response to epinephrine. As well, the investigations used different anesthetics. Given that the locus ceruleus [LC] is differentially affected by barbiturates versus volatile anesthetics, and that the LC mediates cortical flow-metabolism matching via norepinephrine projections, the disparate results may be partially modified by experimental protocol.
Nevertheless, the strength of the evaluation by Mavroudis and colleagues is that 3 different technologies each converge on essentially the same result – that the first 2-3 doses of epinephrine in a porcine cardiac arrest model increase blood flow to both the cortex and subcortical white matter. It is not surprising that they observed ‘diminishing cerebrovascular return’ with additional doses of epinephrine. From a global pressure-flow perspective, the brain is ‘pressure-passive’ at low levels of cerebral perfusion pressure [CPP = mean arterial pressure less intracerebral pressure]. Then over a range of roughly 100 mmHg, increased CPP does not increase cerebral flow because intracerebral arterioles intrinsically vasoconstrict [i.e., the Bayliss effect]. In other words, moving from a CPP of zero to 20 mmHg has a much greater effect on cerebral blood flow than rising from 100 to 120 mmHg. The authors do not report absolute CPP in their manuscript, thus it is hard to know if the reduced effect of subsequent epinephrine doses is associated with auto-regulation.
In summary, the neuro-hemodynamic goal of epinephrine in cardiac arrest is to increase cardiac output and vascular resistance such that mean arterial – and therefore cerebral perfusion – pressure rises without over-constricting the intracerebral vasculature. The commendable investigation of Mavroudis and colleagues reveals that in their porcine model this goal is achieved.