With the beginning of the COVID-19 pandemic, ICUs in the United States and around the world quickly reverted to practices that were common over two decades ago – the administration of deep sedation in mechanically ventilated patients, lack of daily spontaneous awakening and breathing trials, insufficient patient mobilization, and delayed tracheostomies, as well as less communication and restricted access for families. Although understandable in the midst of a national health care crisis, these practices likely contributed to prolonged mechanical ventilation, opioid and benzodiazepine dependance, increased prevalence of coma and delirium, and lingering physical and cognitive disability that continue to be reported in many COVID-19 ICU cohorts (Anesth Analg 2020;131:e40-1). How did COVID-19 rapidly reverse 20 years of progress in evidence-based sedation and ICU liberation?

In patients with COVID-19 ARDS, tachypnea, excessive respiratory drive, and ventilator dyssynchronies are commonly observed and attributed to high dead space (due to pulmonary vascular thromboses), inflammation, and metabolic derangements (e.g., renal and liver failure) (Respir Care 2021;66:1406-15; Crit Care Explor 2020;2:e0208; Intensive Care Med 2020;46:606-18). Difficulties with ensuring low tidal volume ventilation led many clinicians to escalate doses of sedatives and opioids without defined sedation endpoints, resulting in anesthetized, not sedated, patients. For instance, daily doses of propofol in patients with COVID-19 ARDS commonly reached 3,000-4,000 mg according to published reports, roughly a two-fold increase when compared to sedation regimens in ARDS patients before the pandemic (Am J Health Syst Pharm 2021;78:1952-61; Crit Care Med 2021;49:1524-34). Additional sedatives (e.g., midazolam, ketamine, dexmedetomidine) were commonly administered to improve ventilator synchrony, although substantial evidence from the past indicates that deep sedation and coma typically worsen, rather than improve, synchrony (Crit Care Med 2014;42:74-82; J Crit Care 2009;24:74-80). Adjustments in ventilator settings, or even a trial of paralysis, may therefore have been more appropriate interventions as opposed to deepening of sedation to a point that may have compromised cognitive recovery (Intensive Care Med 2020;46:2342-56). High rates of delirium complicate sedation and ventilator weaning in uncooperative or agitated patients. Excessive and prolonged use of sedatives also led to tolerance and subsequent need for prolonged tapers with agents such as methadone, phenobarbital, clonidine, and lorazepam.

Besides the complex physiology of COVID-19 ARDS, concerns over patient safety (prone position), awareness (in paralyzed patients), rapid respiratory decompensation (coughing, auto-PEEP, biting on the ET tube), or the health and safety of staff may have had a role in favoring deep sedation. Staffing shortages, redeployment of less experienced staff in the ICUs, and overwhelming numbers of mechanically ventilated patients further worsened adherence to previously accepted sedation guidelines (The 2018 Pain, Agitation/sedation, Delirium, Immobility [rehabilitation/mobilization], and Sleep [disruption] [PADIS] guidelines).

There are times when deep sedation is in the best interests of the patient. One example is the neurocritical care setting of refractory elevated intracranial pressure (ICP) in traumatic brain injury (TBI) or status epilepticus (Front Neurol 2021;12:750667). The goal in both situations is to suppress any brain electrical activity with intravenous anesthetic therapy, which in turn lowers ICP and suppresses seizure activity. In contrast to ARDS patients under deep sedation, these patients are highly monitored with continuous electroencephalogram (EEG) and often additional multimodal intracranial monitoring in TBI patients, including intracranial pressure monitors, brain tissue oxygen measurements, and cerebral microdialysis. Therefore, the level of the anesthetic can be appropriately and accurately titrated to the desired suppression of brain activity. Doses of sedatives are adjusted continuously based on the patient’s EEG waveform. The goal is always to utilize the minimum dose of the sedative/hypnotic necessary to achieve the therapeutic goal.

Despite the extraordinary cumulative doses of sedatives administered to patients with COVID-19 ARDS, the patients typically have no specific brain monitoring unless they are also treated with neuromuscular blockade (NMB). Even with NMB, continuous monitoring of brain electrical activity with limited channel processed EEG is not routinely performed, and very rarely with a comprehensive montage of EEG electrodes. The limited channel processed EEG as typically used to monitor levels of sedation/anesthesia certainly has limitations. This is especially true if the raw waveform is not appropriately monitored and as the EEG signal deteriorates with poor electrode contact with the forehead surface. In the ICU, there is an extreme fear of awareness when paralytics are used, which leads to excessive infusions of sedative/hypnotic agents. Although the pathophysiology is not well understood, it is well known that patients suffering from COVID-19 can experience a myriad of neurologic complications, including confusion, delirium, encephalopathy, and even frank stroke, which makes the proper titration of sedative/hypnotics even more critical.

Small prospective cohort studies revealed a high prevalence of abnormally suppressed EEG activity (e.g., EEG slowing, discontinuous EEG) in patients with COVID-19 ARDS, suggesting that over-dosing of sedative/hypnotics, or unnecessary administration of sedative/hypnotics in encephalopathic patients with ARDS, may commonly occur (Crit Care Med February 2022). This is concerning given that abnormally suppressed EEG patterns (e.g., burst suppression) have been previously associated with delirium and mortality (Neurocrit Care 2020;33:565-74). The visual and real-time recognition of burst suppression patterns on EEG is relatively straightforward, and their recognition should guide sedation management.

Oversedation is associated with and likely leads to increased days either in coma or delirium, a true cause for alarm. One large multicenter study of adult patients with COVID-19 ARDS reported that 81.6% of patients were comatose for a median of 10.0 days (IQR of 6.0 to 15.0 days), and 54.9% were delirious for a median of 3.0 days (IQR of 2.0 to 6.0 days), for a total of 13 days in either coma or delirium (Lancet Respir Med 2021;9:239-50). During the 21-day study period, patients were only alive without coma or delirium for a median of 5.0 days (IQR of 0.0 to 14.0 days). This is in stark contrast to the MIND-USA 2018 multicenter study in which less than 50% of patients experienced delirium or coma when clinicians adhered to guideline-recommended practices, and the median duration of coma was one day and that of delirium was four days, for a total of five days (N Engl J Med 2018;379:2506-16). These extended periods of pathologic brain states in patients with COVID-19 likely convey a risk of adverse neurocognitive outcomes and impaired survivorship. Patients with prolonged coma and delirium are at risk for not only increased in-hospital complications but post-intensive care syndrome (leading to poor functional outcome), cognitive impairment, and eventual development of dementia (N Engl J Med 2013;369:1306-16; Crit Care Med 2016;44:2003-9).

It is time we start focusing on survivorship, and not only on survival in patients with COVID-19 ARDS and other mechanically ventilated patients. We need to refocus on evidence-based guidelines for sedation management and to utilize the minimum necessary. Such a future has a clear scientific foundation and likely benefit to reduce preventable harm. Increasing the days alive without coma or delirium should be our goal.

Here are three take-home points that we should all seriously consider when sedating patients in the ICU:

  1. When ventilator dyssynchrony is observed in adequately sedated patients, consider adjustments in the ventilator settings or a trial of paralysis instead of further deepening sedation.
  2. Use a limited channel processed EEG to guide therapy with visualization of the raw waveform.
  3. Sedation should not be escalated when the EEG waveform shows burst suppression or is isoelectric.