Cerebral Monitoring and Post-operative Delirium and Outcomes

Overview

Brain monitoring using near-infrared spectroscopy (NIRS) started in 2002 in the operating room of the Montreal heart Institute (MHI). This was followed by the use of somatic NIRS in 2010, transcranial Doppler in 2015 and processed electroencephalogram (pEEG) using Sedline (Masimo, Irvine CA) in 2017. The introduction of those modalities led to significant change in intraoperative management. The goal of these devices is to improve our ability to detect and predict post-operative complications as well as offering insights on how to prevent them. The current project explores in further detail the impact of the introduction of pEEG in the operating room and in the intensive care unit (ICU) on post-operative delirium.

Overview of the problem: Post-operative delirium in cardiac surgery Brain monitoring using NIRS started in 2002 in the operating rooms of the MHI. Since that period, the investigators described our experience on 1613 reported patients from published case reports (n=30), cohort studies (n=1270), randomized controlled trials (n=313), and review articles. In the presence of brain desaturation, the investigators have previously described an algorithmic approach based on a series of interventions (such as head positioning, increasing blood pressure and inspired oxygen etc.) which are triggered by the etiology, one at a time or simultaneously, ideally as early as possible i.e. when the cerebral NIRS value start to decrease by more than 10%. The success rate for correction of brain desaturation was more than 95%. The use of both cerebral NIRS and pEEG led to new insights which were not apparent when these technologies were used in isolation. For instance, brain desaturation resulting from increased pEEG activity will be observed upon awakening. On the other hand, brain desaturation associated with reduced pEEG activity may indicates brain hypoperfusion. A new algorithm to manage cerebral NIRS and pEEG was introduced and published in May 2019. This algorithm has not been validated and could represent a major advance in the care of cardiac surgical patients. One of the potential roles of this type of monitoring would be its ability to predict emergence delirium, postoperative delirium and other end-organ complications. Eventually, the early identification could lead to prevention of post-operative complications such as delirium. Brain oximetry and delirium Between 16% and 30% of elderly hospitalised patients with cardiovascular disease will develop delirium. Delirium and cognitive dysfunction are highly frequent among hospitalized patients across all healthcare settings, but especially in surgical and critical care populations where rates are reported to be as high as 70%. Mortality is increased in patients with delirium after cardiac surgery. Delirium is defined by the DSM-5 as an acute brain dysfunction associated with attention deficit and altered consciousness. Three types of delirium have been described : hypoactive associated with lethargia, hyperactive with agitation and a combination of both that can alternate. There are several predisposing factors such as pre-operative cognitive dysfunction. A recent systematic review has identified the following factors associated with delirium occurrence. Those factors include advanced age, prior psychiatric condition, mild cognitive impairment, and cerebrovascular disease. Post-operative delirium is associated with prolonged duration of hospitalisation, post-operative complications and increased mortality. There have been several reports linking reduced cerebral NIRS measurements, delirium and post-operative cognitive dysfunction among surgical populations. The mechanism through which reduced brain saturation and delirium co-exist could include cerebral malperfusion and venous congestion from right heart dysfunction. Those factors have been recently explored and their association with delirium and post-operative cognitive dysfunction has been reported in the ICU as well as in the operating room. As right heart dysfunction can be associated with prolonged period of venous hypertension, this can lead to interstitial edema which can further impairs end organ perfusion such as the brain. In the context of cardiopulmonary bypass (CPB), this phenomenon might be aggravated by alteration in endothelial glycocalyx resulting in increased permeability of the capillary endothelium. The resulting interstitial edema is thought to be a mediator of adverse outcomes in critically ill fluid overloaded patients. In addition to altered NIRS values observed with delirium after cardiac surgery, abnormal brain electrical activity has been described. Brain electroencephalographic monitoring and delirium Electrical activity of the brain can be monitored through electroencephalography (EEG) and recently bedside monitor of pEEG such as the BIS (Medtronic, USA) or Sedline (Masimo Irvine CA) allow bedside measurement of frontal electrical activity. Delirium is associated with EEG changes, typically global slowing of EEG activity. Significant differences in EEG have been observed in elderly with or without delirium. The alpha, theta and delta wave power and the theta ratio in relation to the other EEG waves are different in patients with or without delirium. Furthermore, significant reduction in EEG activity leading to "burst-suppression" has been related to delirium appearance, prolonged hospital length of stay and mortality. Controversial results have been reported from studies and meta-analysis looking at intervention to correct those factors and improve outcome. However, these observations have never combined information obtained from frontal EEG with brain oximetry. A recent prospective study from Belgium reported an association with post-operative delirium and the information obtained from the combined use of brain NIRS and pEEG. Previous work of delirium and brain oximetry and electroencephalographic monitoring In 2011, the investigators reported our first experience in the use of NIRS in 61 patients undergoing off-pump bypass cardiac surgery as a predictor post-operative cognitive dysfunction. In 2016, an observational study was conducted with 30 consecutive adults with delirium after cardiac surgery. The mean oximetry value decreased from (mean ± SD) 66.4 ± 6.7 to 50.8 ± 6.8 on the first day after delirium onset and increased in patients whose delirium resorbed over the 3 days. The relationship between oximetry, delirium diagnosis, and severity was analyzed with a marginal model and linear mixed models. Cerebral oximetry was related to delirium diagnosis (p≤.0001) and severity (p≤.0001). In that study, somatic oximetry values were normal in patients with delirium. The second study, in 2017, was a case-control retrospective study which included 346 patients, of which 39 (11%), 104 (30%), and 142 (41%) patients presented delirium at 24, 48, and 72 hours post-ICU arrival, respectively. The cumulative fluid balance was associated with delirium occurrence (OR 1.20, 95% CI: 1.066-1.355, p=.003). History of neurological disorder, having both hearing and visual impairment, type of procedure, perioperative cerebral oximetry, mean pulmonary artery pressure pre-CPB, and mean arterial pressure post-CPB also contributed to delirium in the model. In the third study, both a retrospective and prospective cohort were reported including a total of 382 patients. Adult patients who underwent portal venous pulsed-wave Doppler by the attending physician during usual care in the ICU were included in the retrospective cohort (n=237). For the prospective cohort (n=145), patients had a cognitive and echocardiographic evaluation the day before surgery and daily for three days after surgery. Delirium was independently assessed by the nursing staff in the prospective cohort. An association was found between delirium and portal vein pulsatility, a marker of right heart failure, in the retrospective cohort (OR:2.69 CI:1.47-4.90 p=0.001). In the prospective cohort, significant associations were confirmed between the presence of portal vein pulsatility and the development of cognitive dysfunction and asterixis assessed by the investigators (OR:2.10 CI:1.25-3.53 p=0.005 and OR:5.19 \[CI:2.27; 11.88\] p\<0.001) But also the association with delirium detected by the nursing staff was confirmed (HR:2.63 CI:1.13-6.11 p=0.025). Higher NT-pro-BNP measurements (OR:4.03 \[CI: 1.78-9.15\] p=0.001) and cerebral desaturations (OR:2.54 \[CI:1.12-5.76\] p=0.03) were associated with cognitive dysfunction. In this third study, baseline daily pEEG were obtained daily in combination with NIRS values. However, no association was observed between delirium and preliminary waveform analysis. In summary, those studies support an association between episodes of cerebral desaturations measured by brain oximetry and the occurrence of delirium. Venous cerebral congestion as a consequence of right ventricular dysfunction leading to portal hypertension and possibly cytokine release through a cardio-intestinal syndrome appears to be a potential mechanism. As supported by other authors, reduced cerebral NIRS values could be secondary to elevated venous pressure, fluid overload and the latter contribute to a type of congestive delirium. It could also be reflective of increased cerebral metabolism, such as can be observed when patient emerge from general anesthesia after cardiac surgery. However, in those studies, only a single cerebral NIRS and pEEG value were obtained. The combination of both modalities with the use of bilateral cerebral NIRS was not studied prospectively in both the operating room and in the ICU. Currently, it is not known whether the information obtained by these monitors is different in patients presenting an episode of delirium compared to those who do not show any episode of postoperative delirium.