Association of Sedation, Coma, and In-Hospital Mortality in Mechanically Ventilated Patients With Coronavirus Disease 2019–Related Acute Respiratory Distress Syndrome: A Retrospective Cohort Study*

Abstract

OBJECTIVES: 

In patients with coronavirus disease 2019–associated acute respiratory distress syndrome, sedatives and opioids are commonly administered which may lead to increased vulnerability to neurologic dysfunction. We tested the hypothesis that patients with coronavirus disease 2019–associated acute respiratory distress syndrome are at higher risk of in-hospital mortality due to prolonged coma compared with other patients with acute respiratory distress syndrome matched for disease severity.

DESIGN: 

Propensity-matched cohort study.

SETTING: 

Seven ICUs in an academic hospital network, Beth Israel Deaconess Medical Center (Boston, MA).

PATIENTS: 

All mechanically ventilated coronavirus disease 2019 patients between March and May 2020 were identified and matched with patients with acute respiratory distress syndrome of other etiology.

INTERVENTIONS: 

None.

MEASUREMENTS AND MAIN RESULTS: 

Using clinical data obtained from a hospital registry, we matched 114 coronavirus disease 2019 patients to 228 noncoronavirus disease 2019–related acute respiratory distress syndrome patients based on baseline disease severity. Coma was identified using the Richmond Agitation Sedation Scale less than or equal to –3. Multivariable logistic regression and mediation analyses were used to assess the percentage of comatose days, sedative medications used, and the association between coronavirus disease 2019 and in-hospital mortality. In-hospital mortality (48.3% vs 31.6%, adjusted odds ratio, 2.15; 95% CI, 1.34–3.44; p = 0.002), the percentage of comatose days (66.0% ± 31.3% vs 36.0% ± 36.9%, adjusted difference, 29.35; 95% CI, 21.45–37.24; p < 0.001), and the hypnotic agent dose (51.3% vs 17.1% of maximum hypnotic agent dose given in the cohort; p < 0.001) were higher among patients with coronavirus disease 2019. Brain imaging did not show a higher frequency of structural brain lesions in patients with coronavirus disease 2019 (6.1% vs 7.0%; p = 0.76). Hypnotic agent dose was associated with coma (adjusted coefficient, 0.61; 95% CI, 0.45–0.78; p < 0.001) and mediated (p = 0.001) coma. Coma was associated with in-hospital mortality (adjusted odds ratio, 5.84; 95% CI, 3.58–9.58; p < 0.001) and mediated 59% of in-hospital mortality (p < 0.001).

CONCLUSIONS: 

Compared with matched patients with acute respiratory distress syndrome of other etiology, patients with coronavirus disease 2019 received higher doses of hypnotics, which was associated with prolonged coma and higher mortality.

Patients with coronavirus disease 2019 (COVID-19) can have neurologic manifestations, including acute cerebrovascular events and coma (^1–6). In patients with severe COVID-19 requiring admission to the ICU, coma occurs in approximately 15% of patients and is typically diagnosed during the second week of hospital admission (^2,5). Little is known about the etiology and the effect of coma in patients with COVID-19 presenting to an ICU for treatment of acute respiratory distress syndrome (ARDS).

During the COVID-19 pandemic, unprecedented numbers of patients have required sedation in the ICUs and other hospital locations due to prolonged ventilator dependence (⁷). Although sedation is typically required to facilitate mechanical ventilation, deeper sedation levels may be favored by ICU staff to reduce the possibility of a patient’s self-extubation, which carries an increased risk of exposing staff and other patients to COVID-19. Long-acting sedatives were also used more frequently due to impending national shortages of commonly used short-acting sedatives. Finally, clinicians might have favored deep sedation to allow tolerance of lung-protective ventilation strategies.

We tested the hypothesis that patients with COVID-19–associated ARDS compared with patients with ARDS due to other etiologies are at higher risk of increased in-hospital mortality. Contingent on this hypothesis, we also hypothesized that the association of COVID-19 and in-hospital mortality was mediated by a high percentage of comatose days as a consequence of deep sedation during mechanical ventilation.

MATERIALS AND METHODS

Study Design

This retrospective hospital registry study was approved by the Committee on Clinical Investigations at Beth Israel Deaconess Medical Center (BIDMC) in Boston, MA (protocol number 2020P000694) and met criteria for exemption from review. BIDMC is an academic teaching hospital of Harvard Medical School and a tertiary referral center covering intensive care of eight hospitals in the Beth Israel Lahey Health system. During the surge of COVID-19, the medical center doubled its ICU capacity to a total of 140 beds.

Data Sources

We combined multiple electronic data sources to obtain comprehensive information for all included patients. Demographic data, as well as a patients’ past medical history, were collected from online medical records. Laboratory values and ICU data including level of consciousness, ventilator variables, and drug administration were extracted from Metavision, an electronic medical record interface routinely used in all ICUs. Encounter dates and discharge information were retrieved from Casemix and the Admission Discharge Transfer database. Radiographic imaging reports were obtained from the radiology database (supplementary document, section 1, https://links.lww.com/CCM/G322).

Study Population

Adult patients (≥ 18 yr) admitted to the ICUs were included if they were mechanically ventilated and were diagnosed with COVID-19 between March 2020 and May 2020, based on World Health Organization interim guidelines (⁸), or were diagnosed with ARDS between January 2008 and June 2019. Patients with missing baseline characteristics for the propensity matching model were excluded.

Primary Outcomes and Analyses

We tested outcomes in an a priori defined, hierarchical order. The primary outcome was in-hospital mortality. The coprimary outcome was the percentage of comatose days.

The level of consciousness was routinely recorded at least every 4 hours using the Richmond Agitation Sedation Scale (RASS). Patients with a mean RASS score of –3 to –5 were classified as comatose based on previously published literature, irrespective of whether the state was induced by disease or sedation (⁹). The daily sedation variables during the first 10 days of mechanical ventilation are provided in eTable 2 (https://links.lww.com/CCM/G322). We defined the percentage of days spent in coma as the ratio of days with coma during the first 10 days of mechanical ventilation for each patient. The percentage of comatose days was dichotomized into low and high by using the median as the cut off value.

In the primary analysis, we tested the hypothesis that COVID-19 patients had a higher risk of in-hospital mortality compared with non–COVID-19 ARDS patients. Contingent on this assumption, we tested the coprimary hypothesis that the higher percentage of coma was a mediator of in-hospital mortality in COVID-19 patients.

Secondary Outcomes and Analyses

In secondary analyses, we investigated the causes that might affect coma. Sedative and analgesic medications used during mechanical ventilation, the Sedation Burden Index (SBI) during the first 10 days of mechanical ventilation, and structural brain lesions were compared between COVID-19 and non–COVID-19 patients. We tested the hypothesis that the hypnotic agent dose was associated with coma and mediated coma in COVID-19 patients. In addition, we tested the hypothesis that the association between sedation-related coma and mortality may be different than the association between neurologic injury-related coma and mortality.

We collected the daily cumulative doses of sedatives, analgesics, and neuromuscular blocking agents (NMBAs). Opioid doses were converted to oral morphine equivalents, and benzodiazepines were converted to midazolam equivalents for comparison (^10–12). For simplification and better comparison between an individual’s exposure to sedative and analgesic medications, we calculated the SBI which was defined as the cumulative burden from every sedative or analgesic medication that patients received on each day during the 10-day mechanical ventilated period. The average SBI over 10 days was calculated for each patient. To calculate the hypnotic agent dose, we then identified the patient with the highest average SBI over 10 days and expressed individual values as the percent maximum value (supplementary document, section 2, https://links.lww.com/CCM/G322).

All CT scans of the brain and neurologic consultation results were reviewed and confirmed by an experienced intensivist.

Propensity Score Matching

One-to-two propensity score matching for patients with COVID-19 and without COVID-19 was performed as described detailed in the supplementary document, section 3 (https://links.lww.com/CCM/G322) (^13,14). Propensity score estimates, calibrated for both the odd of the exposure (COVID-19–associated ARDS) and the outcome (in-hospital mortality) for each patient, were derived from clinical variables; demographic characteristics (age, sex, and body mass index), comorbidities (Charlson Comorbidity Index), disease severity (Acute Physiology and Chronic Health Evaluation II score), organ failure (renal impairment and severe liver injury), vasopressor support, and baseline ventilator variables at onset of mechanical ventilation (respiratory system compliance, Pao₂:Fio₂ [P/F] ratio, Paco₂, and alveolar–arterial [A-a] gradient) (^15–17). Covariates with residual imbalances following propensity score matching (standardized difference ≥ 0.1) were added as adjustment covariates to a model following matching.

Sensitivity and Exploratory Analyses

We conducted several sensitivity analyses to test the robustness of our findings. In the primary analysis, we matched for possible predictors of the exposure and outcome (¹⁸). In order to address a potential selection bias, we repeated the analysis using the same covariate model in the complete, unselected cohort of all ARDS patients, which are described in detail with other sensitivity analyses in the supplementary document, section 4 (https://links.lww.com/CCM/G322).

With an exploratory intent, the hypnotic agent doses were compared in subgroups of patients who received NMBAs and prone positioning. We examined the trend in ARDS treatment by means of NMBA infusion, prone positioning, and sedative medications used during mechanical ventilation, compared between the COVID-19 pandemic period and the period before the COVID-19 pandemic. Delirium-free days were compared between the two groups (⁹).

Statistical Analysis

Continuous variables and counts are described using mean ± sd or median (interquartile range [IQR]); categorical variables are reported as percentages.

Analyses were performed using chi-square test, multivariable logistic regression, negative binomial regression, and mediation analysis. For mediation analysis, we tested whether COVID-19 patients had a higher percentage of coma during the first 10 days of mechanical ventilation and whether a higher percentage of coma was associated with in-hospital mortality, indicating a possible effect mediation. Conditional on both associations being significant, we used formal mediation analysis to estimate odds ratios of the indirect (mediated) effect of high percentage of comatose state and the total (unmediated) effect of COVID-19 on in-hospital mortality (^19,20). A Cox proportional hazards regression analysis was used to compare the effects of coma on mortality of patients with sedation-related coma versus neurologic injury-related coma. We considered a two-tailed p value of below 0.05 as statistically significant. All analyses were performed using Stata, Version 15.1 (StataCorp LP, College Station, TX).

Power analysis

Based on previously published mortality in mechanically ventilated COVID-19 patients (²¹) and assuming a two-tailed alpha of 0.01 and a power greater than 95% to detect a significant effect, a total sample size of 190 patients was needed.

RESULTS

Study Cohort and Patients’ Cognitive Status Over Time

Among 2,301 patients with a diagnosis of COVID-19 between March 1, and May 12, 2020, 270 patients (11.7 %) were admitted to the ICU. Among those, 141 patients (61.3%) were intubated and mechanically ventilated and therefore included in this study. One-hundred fourteen mechanically ventilated COVID-19 patients (80.9%) were matched with 228 non–COVID-19 ARDS patients (eFig. 1, https://links.lww.com/CCM/G322). Table 1 shows baseline characteristics at the time of initiation of mechanical ventilation and standardized difference after matching. Figure 1 shows the cumulative frequency of death or extubation in each study group over 28 days, as well as the cognitive status during mechanical ventilation for patients in each study group.

TABLE 1. - Baseline Characteristics at the Time of Starting Mechanical Ventilation by Coronavirus Disease Diagnosis; Propensity-Matched Cohort

Characteristics	Non-COVID (n = 228)	COVID (n = 114)	Standardized Difference
Age (yr), mean ± sd	62.3 ± 17.2	60.6 ± 15.9	–0.024
Female sex, n (%)	115 (50.4)	53 (47)	–0.044
Body mass index (kg/m2), median (interquartile range)	29.8 (24.8–36.2)	30.8 (27.3–36.2)	0.011
Acute Physiology and Chronic Health Evaluation II score, median (interquartile range)	25 (20–30)	24 (20–29)	–0.006
Charlson Comorbidity Index, median (interquartile range)	0 (0–2.5)	0 (0–2)	0.027
Respiratory system compliance (mL/cm H2O), mean ± sd	40.7 ± 18.6	39.9 ± 21.8	–0.118
Pao 2:Fio 2 ratio, median (interquartile range)	164 (109–288)	174 (117–270)	0.012
Paco 2 (mm Hg), mean ± sd	48.6 ± 16.3	47.8 ± 13.2	0.017
Alveolar–arterial gradient, median (interquartile range)	435 (256–557)	427 (265–534)	0.040
Creatinine (mg/dL), mean ± sd	1.6 ± 1.3	1.5 ± 1.3	0.019
Total bilirubin (mg/dL), mean ± sd	1.6 ± 1.4	1.7 ± 1.5	–0.288
International normalized ratio, mean ± sd	1.4 ± 3.3	0.7 ± 1.0	–0.105
Vasopressors (µg of norepinephrine equivalents), mean ± sd	0.052 ± 0.177	0.043 ± 0.133	–0.080
pH, median (interquartile range)	7.28 (7.18–7.36)	7.34 (7.27–7.38)	0.149
Minute ventilation (L/min), median (interquartile range)	9.8 (7.8–12.1)	8.7 (7.7–10.0)	–0.357
Glasgow Coma Scale, median (interquartile range)	8 (8–14)	8.5 (8, 15)	0.133
Renal impairment, n (%)	31 (13.6)	15 (13.2)	–0.045
Severe liver injury, n (%)	6 (2.6)	3 (2.6)	–0.090
ARDS severity, n (%)			–0.066
Mild	88 (38.6)	44 (38.6)
Moderate	91 (39.9)	51 (44.7)
Severe	49 (21.5)	19 (16.7)
Cause of ARDS, n (%)
Pulmonary	145 (63.6)	114 (100)	1.068
Documented viral infection	10 (4.4)	114 (100)
Documented bacterial infection	73 (32.0)	0 (0)

ARDS = acute respiratory distress syndrome, COVID = coronavirus disease.

Renal impairment defined as creatinine level ≥ 2.5 mg/dL. Severe liver injury defined as total bilirubin ≥ 5 mg/dL and INR ≥ 1.5. Pulmonary cause of ARDS as opposed to nonpulmonary causes.

Figure 1.: Cumulative frequency of survival and extubation, and patients’ cognitive function during mechanical ventilation in patients with coronavirus disease 2019 (COVID-19) and non–COVID-19-related acute respiratory distress syndrome (ARDS). The cumulative proportion of subjects who died (gray bars) and those who were successfully extubated (white bars) are shown over time. Death substantially differed between patients with and without COVID-19–related ARDS. The shaded areas between the gray and white bars represent the Richmond Agitation Sedation Score (RASS): coma (dark blue, RASS ≤ –3), arousable (blue, –3 < RASS ≤ –1), alert (light blue, –1 < RASS ≤ +1), and agitated (pink, RASS > +1) stage. Patients with COVID-19–related ARDS were more likely to be in a comatose state and had higher in-hospital mortality. ⁺ p < 0.05 for higher in-hospital mortality. ^# p < 0.05 for lower ventilator-free days. ^$ p < 0.05 for higher percentage of coma during the first 10 d of mechanical ventilation.

Primary and Secondary Outcomes

COVID-19 patients experienced a higher percentage of coma during the first 10 days of mechanical ventilation (66.0% ± 31.3%) compared with non–COVID-19 patients (36.0% ± 36.9%). This result was confirmed in an adjusted analysis (adjusted coefficient [aCoef], 29.34; 95% CI, 21.45–37.24; p < 0.001). Mediation analysis revealed that COVID-19 patients experienced a significantly higher percentage of coma, which in turn was significantly associated with in-hospital mortality (adjusted odds ratio [aOR], 5.84; 95% CI, 3.56–9.58; p < 0.001). 58.6% of the effect of COVID-19 on in-hospital mortality was mediated through the indirect effect of coma (aOR, 1.57; 95% CI, 1.22–2.00; p < 0.001). The direct effect of COVID-19 diagnosis on mortality was insignificant (aOR, 1.37; 95% CI, 0.87–2.15; p = 0.17) (eFig. 2, https://links.lww.com/CCM/G322).

Hypnotic Agent Dose and Use of Sedative Medications

The hypnotic agent dose was higher in COVID-19 patients (51.3% ± 20.6% vs 17.1% ± 12.5%) (p < 0.001). The hypnotic agent dose was associated with the percentage of coma (aCoef, 0.61; 95% CI, 0.45–0.78; p < 0.001). Mediation analysis demonstrated that 52.0% of the effect of COVID-19 diagnosis on coma was mediated through the indirect effect of hypnotic agent dose (aOR, 3.08; 95% CI, 1.62–5.83; p = 0.001) (eFig. 3, https://links.lww.com/CCM/G322).

During mechanical ventilation, opioids were the most commonly used medications. There was no significant difference in the numbers of patients receiving opioids between COVID-19 and non–COVID-19 patients (112 [98.2%] vs 214 [93.9%]; p = 0.07). However, the duration of treatment was significantly longer in COVID-19 patients (10.1 ± 5.7 vs 5.6 ± 4.7 d; p < 0.001). COVID-19 patients received a higher proportion of sedative medications and higher doses for a longer time (Table 2) (eTable 3, https://links.lww.com/CCM/G322). The SBI was significantly higher during the first 10 days of mechanical ventilation in COVID-19 patients (p < 0.001) along with a higher percentage of patients who were comatose (Fig. 2).

TABLE 2. - Administration of Sedatives, Analgesics, and Neuromuscular Blocking Agents During Mechanical Ventilation

Drugs Given	Non-COVID (n = 228)			COVID (n = 114)			p a
	No. of Prescriptions, n (%)	Duration ofPrescription (d), Mean ± sd	Dose, Median(InterquartleRange)	No. of Prescriptions, n (%)	Duration of Prescription (d), Mean ± sd	Dose, Median(InterquartleRange)
Opioid (oral morphine equivalents)	214 (93.9)	5.6 ± 4.7	25.1 (12.5–53.1) mg/d	112 (98.2)	10.1 ± 5.7	16.5 (8.1–31.8) mg/d	0.07
Propofol	201 (88.2)	3.9 ± 2.8	1,773 (932–3,176) mg/d	111 (97.4)	7.4 ± 5.0	3,606 (2,188–4,914) mg/d	0.005
Midazolam	120 (52.6)	2.8 ± 2.7	16.1 (3.0–44.8) mg/d	83 (72.8)	6.3 ± 4.4	83.3 (45.5–116.4) mg/d	< 0.001
Lorazepam	42 (18.4)	1.5 ± 1.1	1 (0–2) mg/d	23 (20.2)	1.5 ± 1.1	1.8 (0.5–2.3) mg/d	0.7
Diazepam	1 (0.4)	2	10 (10)	12 (10.5)	3.9 ± 3.0	14.2 (9.8–22.3) mg/d	< 0.001
Dexmedetomidine	103 (44.7)	3.8 ± 2.9	841 (359–1,411) µg/d	65 (57.0)	5.1 ± 3.5	1,157 (806–1,568) µg/d	0.032
Ketamine	2 (0.9)	1.5 ± 0.7	147 (39–254) mg/d	59 (51.8)	7.4 ± 4.5	919 (610–1,570) mg/d	< 0.001
Neuromuscular blocking agent	53 (23.2)	2.6 ± 2.4		64 (56.1)	4.8 ± 3.9		< 0.001
Cisatracurium			138 (89–221) mg/d			182 (112–240) mg/d
Rocuronium			425 (270–579) mg/d			683 (477–845) mg/d

^ap compared between number of treatments.

Figure 2.: Daily mean Pao₂:Fio ₂ (P/F) ratio, percentage of patients in coma and the Sedation Burden Index (SBI). A, Daily mean P/F ratio. B, Daily percentage of patients in a comatose state. Daily percentage of patients in a comatose state defined as Richmond Agitation Sedation Score less than or equal to –3. C, Daily drug burden index. SBI was calculated for an individual by summing the burdens from every sedative medication used on that day, using the equation: SBI =

${\displaystyle \sum }\text{}\frac{\text{D}}{\text{d}+\text{D}}.$

, where D is the drug dose taken, and d is the minimum recommended daily dose. *p < 0.05. COVID-19 = coronavirus disease 2019.

Neurologic Consultation and Brain Imaging

Brain imaging was performed in 36 patients (31.6%) with COVID-19 and 100 patients (43.9%) without COVID-19. Results from neuroimaging reports revealed that patients with and without COVID-19 were not significantly different in terms of pathologic brain lesions (p = 0.76). Seven COVID-19 patients (6.1%) suffered a stroke (3 ischemic and 4 hemorrhagic), and 16 non–COVID-19 patients (7.0%) had pathologic brain lesions (8 ischemic stroke, 4 hemorrhagic stroke, 4 severe cerebral edema). The proportional hazards ratio was not significantly different between sedation-related coma versus coma that was accompanied by an additional neurologic injury (adjusted hazard ratio, 1.34; 95% CI, 0.69–2.59; p = 0.38) (eFig. 4, https://links.lww.com/CCM/G322).

Sensitivity and Exploratory Analyses

The primary results were confirmed throughout all sensitivity analyses. In the complete, unselected cohort of all ARDS patients (n = 3,201) (eTable 1, https://links.lww.com/CCM/G322), COVID-19 patients experienced a higher percentage of coma (aCoef, 35.00; 95% CI, 29.25–40.70; p < 0.001), which in turn was significantly associated with in-hospital mortality (aOR, 3.57; 95% CI, 2.97–4.30; p < 0.001). The effect of COVID-19 on in-hospital mortality was mediated through the indirect effect of coma (p < 0.001). Details of sensitivity analyses are provided in the supplementary document, section 4 (https://links.lww.com/CCM/G322).

Effect Modification and Subgroup Analysis in Patients Who Received NMBAs and Prone Positioning

When considering the effect of adjunctive treatments on the use of sedative medications, an effect modification by prone positioning was observed (p for interaction = 0.04) but not with NMBA use (p for interaction = 0.12). However, a diagnosis of COVID-19 was associated with an increase of hypnotic agent dose in all subgroups of patients (eTable 5, https://links.lww.com/CCM/G322).

Trend in ARDS Treatment

When comparing the common adjunctive therapies used in the management of ARDS prior to and during the COVID-19 pandemic, we found that prone positioning and NMBA use increased during the COVID-19 period (March to May 2020) (48.2% vs 3.5% and 56.1% vs 23.2%, respectively; p < 0.001). Doses and number of sedative and analgesic medications also increased (Fig. 3) (eTable 3, https://links.lww.com/CCM/G322).

Figure 3.: Trends in acute respiratory distress syndrome (ARDS) treatment prior to and during the coronavirus disease 2019 (COVID-19) pandemic. Our COVID-19 treatment period began in March 2020. The average daily sedative medication dose used per patient increased during the COVID-19 pandemic period (*p < 0.001). **p < 0.05 for high medication dose in a patient with COVID-19 compared with a non–COVID-19 ARDS patient.

Delirium-Free Days

Within 28 days, COVID-19 patients had a significantly lower median number of days alive without delirium or coma (delirium-free days) compared with non–COVID-19 patients (median [IQR] = 11 [1–19] vs 21 d [5–25 d]; absolute incidence rate ratio, 0.64; 95% CI, 0.49–0.83; adjusted difference, –6.5 d; p = 0.001).

DISCUSSION

Patients with COVID-19 had a higher risk of in-hospital mortality and experienced a higher percentage of coma during mechanical ventilation compared with patients with ARDS from other etiologies—even after controlling for markers of comorbidities (including markers of heart, renal and liver failure) and ARDS severity (P/F ratio, A-a gradient, Paco₂, pH, and minute ventilation). Coma was identified as a mediator of in-hospital mortality in patients with COVID-19.

COVID-19 is associated with neurologic disturbances and complications including impaired consciousness and stroke (^22,23). In our COVID-19 cohort, clinicians were concerned about a possible nonpharmacologic mechanism of coma in approximately one third of patients to the extent that they requested formal neurologic consultation and brain imaging. Neurologic workup revealed a frequency of stroke confirmed by neuroimaging of 6.1%, which was not different from the 7.0% observed in non–COVID-19 patients.

In our study, patients with COVID-19–associated ARDS received substantially higher doses of hypnotics than patients with ARDS of other etiology, which explained the higher occurrence rate of coma based on our association analysis. Higher utilization of analgesics and hypnotics in mechanically ventilated patients with ARDS due to COVID-19 compared with other etiologies has been reported previously (^24,25). The reasons for the higher requirements are multifactorial: First, based on early treatment suggestions for COVID-19 patients, clinicians opted for prone position therapy more often compared with patients with ARDS of other etiology (48.2 vs 3.5%; p < 0.001). NMBAs were used in most proned patients, which was accompanied by deep sedation targets according to a recommendation (²⁶). NMBA infusions and deep sedation may have been maintained unnecessarily during the periods when the patients were returned to the supine position. We have previously shown that a long duration of deep sedation after termination of neuromuscular blockade is associated with increased mortality (²⁷).

Second, more long-acting sedatives, such as benzodiazepines, were used in COVID-19 patients, which was likely related to national drug shortage reports. On April 10, 2020, propofol and dexmedetomidine were included in the U.S. Food and Drug Administration list of drug shortages (²⁸). In response, our institute issued a document to recommend alternative sedative agents, including benzodiazepines, to be considered during the pandemic drug shortage period. With differences in pharmacokinetic properties such as longer context-sensitive half-life than propofol, benzodiazepines and their metabolites can persist long after infusions are stopped and therefore may cause oversedation (²⁹).

Third, previous publications suggested that patients with COVID-19 can present with different phenotypes, one of which is high ventilatory drive (high dead space and shunt) despite good compliance (^25,30,31). It is possible that clinicians used higher doses of hypnotic agents to protect their patients with COVID-19 from self-inflicted lung injury. This hypothesis is supported by our finding that patients with COVID-19 received higher hypnotic agent doses and paralytics more frequently, resulting in lower minute ventilation compared with patients with ARDS of other etiology. Finally, in order to mitigate the risk of viral spreading across healthcare providers and patients, strict infection control protocols, including isolation of patients, personal protective equipment, and respiratory protection, were recommended (³²). Deep and prolonged sedation may have been unintended consequences. Maintaining deep sedation may also be anxiety driven; ICU staff may have been focused on the prevention of patients’ self-extubation more for patients who tested positive for COVID-19.

Mediation analysis demonstrated a strong relationship between coma and increased in-hospital mortality. Expectedly, we found that a longer period of sedation was associated with a higher occurrence rate of delirium, which could have contributed to the effects of sedation on mortality. Our findings are supported by previous studies demonstrating that deep sedation or coma lead to immobilization, delirium, delays in extubation, and delays in ICU discharge (^33–35). Such delays increase the risk of nosocomial infections and other complications that can contribute to mortality (³⁶). Current sedation guidelines and sedation bundles should be considered and applied when possible (^37,38).

Our study is the largest cohort study investigating sedative drug usage and cognitive function in mechanically ventilated COVID-19 patients. We also have a highly detailed and granular database of non-COVID ARDS patients, with which we were able to compare characteristics, clinical practice, and outcomes.

However, several limitations arise from the retrospective design of our study. First, although we matched patients based on most relevant factors such as heart, liver, and renal failure at admission, as well as severity of ARDS, we cannot exclude the relevance of unmeasured confounders. Second, to the best of our knowledge, there is no robust way to combine the effects of different drug compounds. We adapted the drug burden index as a SBI to compare the effect of summative analgesic and sedative doses in critically ill patients (^39–41). We cannot identify the cause of high sedative medication needs, even though we controlled for all available variables that affect respiratory drive (P/F ratio, A-a gradient, Paco₂, pH, and minute ventilation). Finally, it is possible that severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) may have had direct effects on the brain which could not be identified by CT imaging. Leukoencephalopathy is a rare CNS complication in SARS-CoV-2 which cannot be diagnosed with a CT (⁴²), and it is possible that patients who were identified as not having a structural neurologic disease had an unidentified leukoencephalopathy.

CONCLUSIONS

Patients with COVID-19 received higher doses of hypnotics during mechanical ventilation, which was associated with prolonged coma and increased in-hospital mortality. Clinical practices focusing on safely minimizing analgesic and sedative doses and duration of administration should be applied.

ACKNOWLEDGMENTS

We are grateful to Dr. Timothy T. Houle for providing us with an independent statistical advice during the per-review process.