Do Our Sedation Practices Contribute to Increased Mortality in Coronavirus Disease 2019–Related Acute Respiratory Distress Syndrome?*

Abstract

Coma, a clinical manifestation of acute encephalopathy (1), and most often defined in the ICU by a Richmond Agitation-Sedation Scale (RASS) score of –4 or –5, characterizes the patient who is in a state of unresponsiveness, with eyes closed, and who does not arouse to even vigorous stimulation (2). An underlying medical condition (e.g., acute neurologic injury) and sedative therapy are the two most common causes for coma in the ICU. Medication-induced coma is associated with increased mortality, a prolonged duration of ventilation and ICU stay, and increased delirium (3–6). Since early 2020, nearly 1 million adults worldwide have been admitted to an ICU with severe acute respiratory coronavirus 2 (SARS-CoV-2) and required mechanical ventilation for management of viral pneumonia and acute respiratory distress syndrome (ARDS) (7). These patients are at high risk for neurologic and neuropsychiatric complications both during and after their ICU stay due to high systemic inflammation and neuroinflammation, the effects of multiple organ system failures, and the administration of deep sedation during prolonged mechanical ventilation (8). Although SARS-CoV-2 is generally not detected in the cerebral spinal fluid, the role of parainfectious neurologic disease and adaptive immune responses to severe coronavirus disease 2019 (COVID-19) on the brain may contribute to symptoms of brain dysfunction (9,10). In a recent cohort study of 2,088 patients admitted with ARDS secondary to SARS-CoV-2 to 69 ICUs across 14 countries, 82% experienced coma (average duration 10 d), and 55% developed delirium (average duration 3 d) (11). Most patients received continuous sedative infusions: benzodiazepines and propofol were administered to 64% and 71%, respectively, for a median of 7 days. Although benzodiazepine use and a lack of family visitation were identified as modifiable risk factors for delirium, risk factors for coma were not investigated. The very high prevalence of coma during the first spring 2020 COVID-19 surge may be related to knowledge gaps regarding optimal ventilator strategies and sedation techniques and the abandonment of light sedation practices in the face of clinician fear and shortages of ICU-trained providers, personal protective equipment, and propofol (12). In this issue of Critical Care Medicine, Wongtangman et al (13) provide us further insight into the contributory role of coma on the outcomes of patients with COVID-19 ARDS. The authors hypothesized adults with COVID-19–associated ARDS are at higher risk of in-hospital mortality (compared with disease severity-matched adults with ARDS from other causes), and this mortality may be mediated by prolonged coma (defined as a RASS ≤ –3) from either neurologic injury or sedating medication. In their retrospective, single-center, 3-month cohort analysis, 114 COVID-19 ARDS patients were matched with 228 non–COVID-19 ARDS patients using baseline disease severity. The propensity of in-hospital mortality was derived from demographic characteristics, comorbidities, disease severity, organ failure, vasopressor support, and ventilator variables at mechanical ventilation onset. After converting all sedatives to daily midazolam equivalents and opioids to daily morphine equivalents, the cumulative exposure to daily sedative and analgesic medications over the course of the first 10-day mechanical ventilation period was presented as a sedation burden index (SBI). Using the patient with the highest SBI in each group, each patient’s SBI was presented as the percent maximal SBI value. Over the first 10 days of mechanical ventilation, the COVID-19 patients were found to experience a higher percentage of coma than non–COVID-19 patients (66 vs 36%; p < 0.001). Mediation analysis revealed coma in the COVID-19 (vs non–COVID-19) patients to be associated with a nearly six times greater in-hospital mortality (adjusted odds ratio [aOR], 5.84; 95% CI, 3.56–9.58; p < 0.001), whereas the direct effect of a COVID-19 diagnosis itself on mortality was insignificant (aOR, 1.37; 95% CI, 0.87–2.15; p = 0.17). The SBI, three times higher in the COVID-19 patients, was independently associated with coma, and 52% of the effect of the COVID-19 diagnosis on coma was mediated through the indirect effect of the SBI (aOR, 3.08; 95% CI, 1.62–5.83; p = 0.001). Adjustment for factors including continuous neuromuscular blocker use or proning did not influence these results. The proportional hazards ratio (HR) for in-hospital mortality was not significantly different between sedation-related coma versus coma that was accompanied by an additional neurologic injury (adjusted HR, 1.34; 95% CI, 0.69–2.59; p = 0.38). Within 28 days, COVID-19 patients had a significantly lower median number of days alive without delirium or coma compared with non–COVID-19 patients (adjusted difference –6.5 d; p = 0.001). As the authors note, the results of their analysis may be limited by the inability to account for unmeasured confounders—a common challenge in all mediation analysis (14). Its retrospective nature precludes knowing the clinical rationale for the interventions (e.g., sedatives) administered and whether targeted goals were reached. Only one third of patients underwent neurologic screening and imaging; post-hospital outcomes were not considered. Results from a single center might not be applicable to other centers; patient data from the first Spring 2020 ICU surge may be less applicable 1 year later given our knowledge of severe COVID-19 ARDS has improved (15). Despite these potential limitations, the results of the rigorous analysis by Wongtangman et al (13) provides important take home messages to both clinicians and researchers. Apart from raising awareness about deep sedation practices and resultant prolonged coma in patients with COVID-19 ARDS, their article (13) is the first ICU report using mediation analysis techniques to demonstrate a relationship between sedative administration, coma occurrence, and mortality in the ARDS population. Mediation analysis offers advantages over traditional time-dependent multivariable analysis given that it seeks to further understand the mechanism(s) by which sedation may impact mortality (14). The traditional use of confounders may have falsely suggested that sedation and coma solely contribute to increased mortality in this population. By performing mediation analysis, we now understand that COVID-19 ARDS leads to increased sedative use that “both” contributes to increased coma and also to increased mortality. In contrast to a 15-year-old report from Ouimet et al (16), where coma related to neurologic injury was not associated with increased mortality, the authors report coma in the ICU is associated with increased mortality regardless of whether coma is medication induced or the result of neurologic injury. Although a tremendous amount of research is focused on the acute neurologic sequelae of severe COVID-19 infection, this report suggests that acute neurologic events like stroke are not higher when COVID-19 is present, yet iatrogenic sedative-associated coma is rampant and associated with worse outcomes. The high doses of sedatives used in adults with COVID-19 ARDS, and its association with prolonged coma and increased mortality, highlights the importance of employing the Assessment, Prevention, and Manage Pain; Both Spontaneous Awakening Trials and Spontaneous Breathing Trials; Choice of Analgesia and Sedation; Delirium: Assess, Prevent, and Manage, Early Mobility and Exercise, and Family Engagement and Empowerment (ABCDEF) bundle (i.e., “Bundle”) in this population (17). The Bundle strives to optimize pain management, avoid deep sedation, reduce delirium, shorten the duration of mechanical ventilation, minimize ICU-acquired weakness, and foster ICU patient and family involvement in care processes. Multiple studies demonstrate impact of Bundle use on ICU patient care, outcomes, and healthcare costs in non–COVID-19 critically ill adults (17–19). Although COVID-19–related issues and barriers may reduce Bundle use, particularly at centers experiencing a surge of patients (12), ignorance of the Bundle will influence both short- and long-term patient outcomes (3). It is abundantly clear that light sedation is not possible during proning or neuromuscular blockade use—but once these interventions are no longer required, a rapid transition to light sedation should be undertaken. The lack of such a mindset was apparent in the international cohort study where assessment of pain (73%), sedation-agitation (98%), and delirium 83%) was high (on eligible ICU days), but a spontaneous awakening trial was performed on only 24% of eligible days, a spontaneous breathing trial on 23%, and early mobility and exercise on 34%. The recently published Maximizing the Efficacy of Sedation and Reducing Neurological Dysfunction and Mortality in Septic Patients With Acute Respiratory Failure (MENDS2) trial attests to the fact that mechanically ventilated septic patients, many of them with ARDS—similar to COVID-19 ARDS patients—can be safely managed with light sedation and that outcomes, including cognition at 6 months, are similar regardless of whether propofol or dexmedetomidine is used (20). Clinicians therefore have the choice of two nonbenzodiazepine agents to use to target light sedation in their COVID-19 ARDS patients. Reemploying the use of evidence-based strategies developed over the past 20 years through rigorous controlled trials is one of the best mechanisms by which to help critically ill adults with COVID-19 liberate from the ICU and transition toward recovery and survivorship (3,4). Although adults with severe ARDS have been underrepresented in these investigations and may require periods of deep sedation and neuromuscular blocker therapy during their ICU course, an analgesia-first and minimization of sedation approach to therapy should be promoted (4,20). In the setting of lightened sedation, distress-related symptoms (e.g., pain, anxiety, and dyspnea) and patient-ventilator asynchrony should be assessed and managed by the interprofessional ICU team (21).

Topics & Concepts