Managing the cerebral complications of acute respiratory distress syndrome

Acute respiratory distress syndrome (ARDS) comprises 10–30% of critical care admissions worldwide [1]. ARDS outcomes have improved secondary to ventilatory strategies that may potentially mitigate ventilator-induced lung injury [2]. Interventions such as low tidal volume ventilation, minimization of the fraction of inspired oxygen (FiO₂), positive end-expiratory pressure titration, neuromuscular blockade, prone positioning, and facilitating “lung protection” with veno-venous extracorporeal membrane oxygenation (V–V ECMO) may have collectively improved mortality rates [3]. As the proportion of ARDS survivors has increased, recognition of the neurological sequelae in ARDS survivors has gained traction. [4]

Manifestations of ARDS-associated cerebral complications are heterogeneous and comprised of entities, such as ischemic stroke, hypoxic-ischemic brain injury, intracranial hemorrhage, seizures and cerebral edema with increased intracranial pressure (ICP), as well as neurocognitive dysfunction and psychological alterations [5]. The incidence of post-ARDS neurological sequelae ranges from 50% to 80% [6]. The pathophysiology of these disease processes is poorly understood in the context of ARDS but may be reflective of within patient individual pathology and exacerbated by clinical interventions that facilitate lung protective ventilation [6].

The conundrum of ARDS management is the recognition that implementation of lung protective ventilation strategies may exert a deleterious effect on cerebral physiology [7]. In other words, “saving the lung may harm the brain”. Clinical contributors to the pathophysiology of ARDS-associated cerebral complications can be approached on the basis of three physiologic variables: (a) arterial carbon dioxide tension (PaCO₂); (b) arterial oxygen tension (PaO₂); and (c) decreased cerebral venous return [6]. First, hypercapnia is common in ARDS and elevated PaCO₂ leads to cerebrovascular vasodilation and cerebral hyperemia with increased ICP [6, 8]. Conversely, an abrupt PaCO₂ reduction following initiation of V–V ECMO is associated with cerebral complications, which may stem from cerebral vasoconstriction with downstream ischemic stroke or hypoxic–ischemic brain injury [9]. Second, minimization of FiO₂ is routine to mitigate hyperoxia-induced lung injury and adsorption atelectasis [10]. This strategy may expose ARDS patients to severe reductions in PaO₂ (< 60 mmHg) which could render the brain vulnerable to brain tissue hypoxia and neuronal injury. Third, prone positioning improves gas exchange, is associated with decreased ventilation duration and mortality [10]. However, the risks of prone positioning include reducing jugular venous return from the intracranial compartment (due to zero-degree bed positioning and increased intra-thoracic pressure) and consequent increased ICP [11]. Similarly, high intra-thoracic pressure related to positive end-expiratory pressure or plateau pressure can further reduce venous return and aggravate ICP [12]. In summary, the lung protective strategies may negatively impact cerebral physiology and subsequently lead to neurological complications in ARDS patients.

Nullifying the downstream neurological sequelae of ARDS after which they occur is challenging. As such, a proactive approach can be taken that focusses on early assessment of individual patient cerebral physiology to mitigate cerebral complications before they occur. This approach requires expeditious and timely evaluation of within patient cerebral physiology using early neuroimaging and non-invasive bedside neuromonitoring. Doing so can provide clinicians with the requisite physiologic and clinical data points to individualize the implementation of interventions designed to improve pulmonary function while minimizing the risk of cerebral complications. For example, in an ARDS patient with evidence of increased ICP, implementation of prone positioning or permissive hypercapnia may be avoided at the expense of potential deleterious effects on cerebral physiology. Additionally, in an ARDS patient with evidence of ongoing cerebral ischemia, tolerance of severe hypoxemia may be avoided to ensure the vulnerable brain has adequate oxygen delivery to sustain cellular and structural integrity.

In these complex clinical scenarios of competing interests, it is imperative that clinicians incorporate expeditious diagnosis and recognition of individual patient pathophysiology with available diagnostic modalities (Fig. 1). Conventional evaluation of the brain is difficult in ARDS patients due to the confounding of the clinical examination with intravenous sedatives and paralytics that are used to facilitate safe ventilation strategies. Further, routine neuroimaging with computed tomography or magnetic resonance imaging is challenging due to potentially precarious patient transfers. As such, increased reliance upon alternative diagnostics is increasingly being recognized as a useful tool to glean insights into patient-specific pathophysiology and guide individualized management decisions.

Fig. 1.

Lung protective ventilation strategy and its impact on cerebral physiology in acute respiratory distress syndrome (ARDS). ARDS management includes lung protective ventilation, permissive hypercapnia, relative hypoxemia, and increased intra-thoracic pressure from high positive end-expiratory pressure, each contributing to increased intracranial pressure (ICP) and secondary brain injury. The bundle approach incorporating bedside non-invasive neuromonitoring tools can facilitate early detection of cerebral complications and early and timely management strategies. Non-invasive techniques such as transcranial Doppler (TCD) may provide insights into the degree dysfunctional cerebral autoregulation and potentially assist in identifying patient-specific perfusion thresholds in ARDS patients. Electroencephalography can be utilized to diagnose non-convulsive seizures or localize cerebral pathologies related to ARDS complications

In particular, point-of-care diagnostic tools can inform within patient physiology that is reflective of intracranial pressure and cerebral blood flow. Such crucial clinical physiologic considerations can enable clinicians to tailor decision-making. Non-invasive estimation of ICP can be gleaned with use of optic nerve sheath diameter ultrasonography [13], quantitative pupillometry [14], and transcranial Doppler (TCD) [13]. In each case, links between intracranial hypertension and elevated optic nerve sheath diameter on ultrasound (> 6 mm), reduced neurological pupil index, and non-invasive ICP quantification with TCD have been evaluated in critically ill patients [13–15]. Tools such as these may be used as screening measures for intracranial hypertension. If increased ICP is present with the aforementioned bedside tools, clinicians may opt to consider alternate management modalities that do not severely exacerbate intracranial hypertension such as extreme permissive hypercapnia or prolonged prone positioning, and instead utilize V–V ECMO to facilitate improvements in arterial oxygenation and avoidance of hypercapnia.

Middle cerebral artery flow velocity (MCA_FV) can be used as a surrogate for cerebral perfusion in critically ill patients. MCA_FV assessment may be used to personalize decision-making in circumstances whereby patients are exhibiting signs of cerebral ischemia. For example, in an ARDS patient with signs/symptoms of diffuse or focal cerebral ischemia, MCA_FV can guide titration of cardiovascular hemodynamics and careful manipulation of PaCO₂ (especially after V–V ECMO initiation). Cavayas et al. conducted an retrospective analysis of > 11,000 V–V ECMO patients and found that the rate of cerebral complications was higher in patients in whom the PaCO₂ decreased by > 50% following V–V ECMO than those without [9]. Although these data need further validation, cautious PaCO₂ reduction in severe ARDS patients undergoing V–V ECMO may be considered to limit the incidence of cerebral complications.

Individualized clinical decisions to mitigate ARDS-associated cerebral complications must integrate neuromonitoring-guided evaluation and selection of lung protective interventions with consideration of the individual patient cerebral physiology. Such individualization of management decisions could prevent additional neurological injury, which may improve neurological outcomes in ARDS survivors by reducing cerebral complications and potentially have an impact on post-intensive care syndrome outcomes.