It All Circles Back to Cerebral Autoregulation: Understanding the Risk of Hypocapnia and Arterial Hypertension When Initialing Pediatric Extracorporeal Membrane Oxygenation*

Clinicians caring for children with cardiopulmonary failure using extracorporeal membrane oxygenation (ECMO) are refining bedside clinical practices by focusing on minimizing morbidity associated with neurologic complications. Neurologic complications are the main drivers of survival and survivorship (1). Both enhanced neuromonitoring (2,3) and neuroimaging (4) serve to characterize neurologic morbidity to help the field establish specific objectives for improvement. In this issue of Pediatric Critical Care Medicine, Shah et al (5) investigate potentially modifiable factors known to have a strong physiological foundation. The study investigates the relationship between relative changes in Paco2 and or mean arterial blood pressure early after starting ECMO and the risk of neurologic complications. The investigators use the registry of the Extracorporeal Life Support Organization (ELSO) and define neurologic complications as any report of seizures, CNS infarction or hemorrhage, or brain death. Among greater than 7,000 ECMO runs, 15% were reported to have neurologic complications. The results suggest an association between a greater magnitude of decrease in Paco2 and arterial hypertension following ECMO initiation with neurologic complications; in addition, among the group with a greater magnitude change in Paco2, there was an increased odds of neurologic complication per percentile increase in arterial blood pressure. This analysis of the ELSO registry is very important as it adds epidemiological evidence to previously published reports suggesting an increase in odds of death with decreases in carbon dioxide in children (6) and an increase in odds of neurologic adverse events in adults supported with venovenous ECMO (7,8). The conclusions are consistent with the physiological evidence suggesting alterations in cerebral vascular reactivity and blood flow with changes in blood pressure published in venoarterial animal models of ECMO (9,10) and in infants with ECMO (11). Decades ago, the fields of physiology and anesthesiology published experiments that lay the foundation for our understanding of cerebral autoregulation in rats (12), cats (13), primates (14,15), and humans (16). Collectively, there is biological plausibility to the findings generated with the analysis by Shah et al (5): carbon dioxide is one of the most potent and predictable modifiers of cerebral blood flow (and blood volume) and acute arterial hypertension above the upper limit of cerebral autoregulatory increases cerebral blood flow. A unit change in Paco2 between 20 and 60 mm Hg changes cerebral blood flow by 2%–3%; cerebral blood flow changes induced by Paco2 start within 2 minutes and reach equilibrium within 12 minutes (13), and prolonged sustained changes in Paco2 lead to adaptation after which 36 hours, cerebral blood flow will have returned to baseline (17). Additionally, ischemic areas of the brain may exhibit altered regional carbon dioxide reactivity (15). Arterial hypertension in the context of cerebral ischemia has been shown to increase cerebral blood flow (14) which may add to brain tissue injury. The analysis by Shah et al (5) suggests that arterial hypertension is associated with a higher probability of neurologic complications and that it adds to the association with hypocapnia. Faced with these conclusions, it is important for the reader to understand how rapid reduction in Paco2 or arterial hypertension may occur when initiating ECMO despite close frontline clinical attention. First, the reduction in Paco2 results from blood draining rapidly from the child and circulating through an efficient membrane lung. The membrane is connected to a source of gas regulated by a flow meter adjusted manually. On initiation after connecting the ECMO tubing to the cannulas, the clinician responsible for the ECMO system must carefully dial three parameters: 1) the pump’s blood flow—which is adjusted to support the targeted cardiac output; 2) the gas fraction of oxygen to the air delivered to the membrane lung (Fdo2); and 3) the gas flow in L/min (often called sweep) adjusted to target the ventilation of the membrane lung. Carbon dioxide diffusion is very efficiently achieved by most contemporary membranes lungs available on the market for ECMO in different sizes. Generally, in situations where the ECMO pump blood flow provides the majority of cardiopulmonary support, either in venoarterial or venovenous modes, on initiation clinicians assume that the gas flow delivered to the membrane will be the main modifier of the Paco2. Most ECMO systems used in intensive care settings do not deliver carbon dioxide like in cardiopulmonary bypass systems. Hence to precisely adjust the change in Paco2 parameter, the clinician must dial a gas flow meter to the estimated value that will “ventilate the membrane” while not leaving the circulating blood precipitously hypocapnic, and then titrate the “ventilation of the membrane” based on the measurement of patient arterial gases. Like a pilot responsible for an airplane’s lift-off, the clinician responsible for the ECMO system must ensure precise adjustments to perfuse cerebral and systemic circulations. However, differently to an airplane cockpit’s monitors, real-time feedback monitoring of Paco2 is very limited: raw values of end-tidal capnography are reduced given the reduction in blood flow through the pulmonary circulation, transcutaneous capnometry may or may not be reliable depending on the patient’s impaired perfusion or hemodilution (i.e., when ECMO is initiated with crystalloids). Cerebral oximetry (% Sco2) estimated from near-infrared spectroscopy prefrontal probes do not require a pulsatile signal and may offer some indication of tissue oximetry and relative decrease in perfusion associated with a decrease in cerebral blood volume and blood flow with hypocapnia. However, Sco2 is a parameter that also increases with an increase in patients’ systemic oxygenation following ECMO initiation and does not only reflect cerebral perfusion changes induced by Paco2. Ultimately, frequent early manual sampling and analysis of patient arterial blood gases are the most precise means of bedside monitoring changes in carbon dioxide available in intensive care settings. In situations when ECMO is initiated urgently in the absence of intra-arterial blood monitoring, sampling from the ECMO circuit may offer an alternative to direct patient sampling. Clinicians responsible for ECMO are expected to have a thorough understanding of cerebral autoregulation and the study by Shah et al (5) helps to operationalize safety boundaries for frontline clinicians. There are limited studies conducted in the pediatric ECMO population that guide clinical practice on how to target the rate of carbon dioxide change, that is, 5 or 10 mm Hg per 12-hour or more or less? Moreover, there are no studies in ECMO patients that guide when a prolonged duration of hypercapnia before cannulation should alter clinicians’ approach. For example, preclinical physiology principles would suggest that a patient with acute respiratory distress syndrome who has been hypercapnic for 10 days should be given more time to normalize their Paco2 than a patient with acute respiratory acidosis from airway obstruction from aspiration of a foreign body? Management of ECMO patients is extrapolated from principles in basic physiology and from other disease states like stroke or trauma or cardiopulmonary bypass literature. Second, if we examine how hypertension may occur, arterial hypertension following initiation of ECMO may be the result of multiple patient-related and ECMO system-related factors. On initiation, the pump’s blood flow is dialed to achieve a starting target flow prescribed set to replace the cardiac output needed. This starting pump flow is based on cardiac output estimates delineated by patient size (e.g., 100 mL/kg for biventricular physiology) and physiology on the assumed increased needs (e.g., 150 mL/kg for single ventricle physiology or more in severe septic shock). Age-based blood pressure targets are then applied to immediately adjust pump flows. At this time of ECMO initiation, unless the clinician deliberately reflects on the age-appropriate normative arterial blood pressure targets, arterial hypertension may occur from the pump being dialed up concurrently to pharmacological therapies already in place. Following this, it is not unexpected to assume that acute arterial hypertension above the cerebral autoregulatory threshold could occur and increase cerebral blood flow. In animal models, this would add to the magnitude of ischemic injury. In the study by Shah et al (5), the authors provide evidence that arterial hypertension above the normative mean arterial blood pressure for age increases the risk of neurologic complications. Shah et al (5) thoughtfully considered different potential confounders of neurologic complications such as the context of cardiopulmonary arrest and ECPR or the use of carotid cannulations. Although there remains confounding indication intrinsic to the study’s design, they still found that a significant decrease in carbon dioxide and acute arterial hypertension was associated with neurologic complications. Some other parameters that may alter cerebral blood flow or the risk of neurologic complications that were not evaluated are hyperoxemia following cannulation (8,18,19), changes in hematocrit with hemodilution, rates of rewarming, and disease states with increased endothelial injury and inflammation (20). Given that the neurovascular unit and blood–brain barrier may not be equally modified across different groups of children with ECMO, important studies are ongoing to better understand these settings (21,22). Although we recognize the intrinsic limitations of a registry-based retrospective study, these analyses provide motivation for increased monitoring and for the study of personalized-patient safety boundaries. Shah et al (5) also illustrate that optimizing outcomes will result from refining the bedside clinical physiological management and providing a more comprehensive “cockpit” to clinicians initiating ECMO. This begins with ensuring that physiologic explicit deliberate targets are set before initiation and formalized in order sets or reproducible algorithms for the clinician, similar to a “flight plan,” for example, Paco2 rate change targets, and mean intra-arterial blood pressure targets. It also suggests that there may be an opportunity for improvement by motivating the industry to design sensors (23) to measure arterial carbon dioxide and test ECMO systems with closed-loop feedback systems. Until then, it seems simply logical and necessary to ensure ECMO systems used in infants and children include equipment with membranes tailored to their size and physiology, to deliver and adjust precise gas and blood flows (e.g., microflow meters).

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.