Intensive care unit delirium in patients with severe COVID-19: A prospective observational cohort study.

Background: Delirium is common in patients with severe coronavirus disease-19 (COVID-19). The purpose of our study was to determine whether severe COVID-19 is an independent risk factor for the development of delirium in patients treated in the intensive care unit (ICU).
Methods: This prospective observational cohort study involved 162 critically ill patients admitted to a multidisciplinary ICU during 2019 and 2020. A validated screening tool was used to diagnose delirium. Multiple delirium risk factors were collected daily including clinical characteristics, hospital course, lab values, vital signs, surgical exposure, drug exposure, and COVID-19 characteristics. After univariate analysis, a multivariate logistic regression analysis was performed to determine independent risk factors associated with the development of delirium.
Results: In our study population, 50 (31%) patients developed delirium. A total of 39 (24.1%) tested positive for COVID-19. Initial analysis showed COVID-19 to be more prevalent in those patients that developed delirium (40% vs. 17%; P = 0.003). Multivariate analysis showed opioid use (odds ratio [OR]: 24 [95% confidence intervals (CI): 16-27]; P ≤ 0.001), benzodiazepine use (OR: 23 [95% CI: 16-63] P = 0.001), and estimated mortality based on acute physiology and chronic health evaluation IV score (OR: 1.04 [95% CI: 1.01-1.07] P = 0.002) to be independently associated with delirium development. COVID-19 (OR: 1.44 [95% CI: 0.13-10.6]; P = 0.7) was not found to be associated with delirium.
Conclusion: Delirium is prevalent in critically ill patients admitted to the ICU, including those with COVID-19. However, after adjustment for important covariates, we found in this cohort that COVID-19 was not an independent risk factor for delirium. Copyright:

INTRODUCTION

Coronavirus disease-19 (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is notorious because of its impact on the respiratory system.[1234] Intensive care is often required in these patients due to the severe respiratory failure that can result from COVID-19;[4] approximately one-quarter of those hospitalized require treatment in the intensive care unit (ICU) due to refractory hypoxemia, shock, or multiple organ failure.[5] However, as we approach the second anniversary of the global pandemic, evidence continues to emerge demonstrating that the effect of this virus reaches far beyond the lungs. Due to the broad tropism of SARS-CoV-2, such effects include gastrointestinal, renal, hepatic, cardiac, and neurologic manifestations.[123467891011121314]

Delirium, which is associated with disturbances in consciousness, attention, perception, and behavior, is an organic brain syndrome with a fluctuating course.[691415] It is an independent predictor of mortality, increased length of stay, ventilator utilization, and lasting cognitive impairment.[12467910111213] Delirium has long been considered an indicator of systemic critical illness.[791113] In COVID-19, the World Health Organization has recognized delirium as a core symptom of SARS-CoV-2 infection;[16] interestingly, delirium may be the only finding in those without overt respiratory symptoms.[1261014] Furthermore, a recent meta-analysis found that delirium is associated with and has the potential to predict, which patients will experience higher morbidity and mortality as a result of COVID-19.[15]

Early studies from Wuhan describe neurologic symptoms of SARS-CoV-2 infection; “impaired consciousness” was seen in 7.5% of patients in one of the first retrospective reviews reporting central nervous system (CNS) dysfunction.[7816] However, such cases were likely underreported as a validated delirium screening tool was not utilized.[11] Recent studies have suggested that delirium commonly occurs in COVID-19 patients with frequency increasing in relation to the severity of illness.[12713171819202122] The prevalence of delirium has been reported to range from 65% to 79.5% in those COVID-19 patients admitted to the ICU.[12111922] However, delirium is prevalent in those requiring intensive care for any reason.[478913]

Although it has been postulated that delirium occurs in greater frequency than would otherwise be expected in COVID-19 patients,[71217] whether SARS-CoV-2 infection is an independent risk factor for delirium has not been clearly established. Additional factors that contribute to the development of delirium in COVID-19 patients include those associated with pneumonia and ARDS, such as hypoxia and respiratory acidosis.[12681923] ICU admission, the administration of deliriogenic drugs, and the social isolation that results from quarantine further increase the likelihood of developing delirium in these susceptible patients.[1467911141924]

In a prospective cohort of consecutively admitted patients to a multidisciplinary ICU, we aimed to investigate whether SARS-CoV-2 infection is an independent risk factor for developing delirium in those patients suffering from severe COVID-19.

METHODS

Study population

This prospective observational cohort study was conducted from May 1, 2019 to October 31, 2020 at a 30-bed multidisciplinary ICU in Mayo Clinic Hospital, Phoenix, Arizona. This ICU is fully staffed by intensivists 24/7. Approval was provided by the Mayo Institutional Review Board (IRB) before initiation of data collection; the need for patient consent was waived. This study was performed in accordance with the Helsinki Declaration pertaining to medical research involving human subjects.

Daily ICU admissions were screened by the investigators for the assessment of inclusion and exclusion criteria. Inclusion criteria: consecutive critically ill patients ≥18 years of age admitted to the ICU during the study period; for COVID-19 patients, a COVID-19 polymerase chain reaction (PCR) test positive within 14 days. Exclusion criteria include patients whose code status is either do no resuscitate or do not intubate (DNR/DNI) and comfort care patients, presence of delirium at the time of ICU admission, and patients who had not agreed to the use of their medical records for research. Furthermore, those patients with a history of substance abuse, psychiatric disease, cognitive impairment, incarcerated persons, and pregnant persons were excluded as vulnerable populations in accordance with the IRB this study was performed under. When patients required readmission to the ICU after discharge, only data from the first admission were analyzed.

Data collection

Comprehensive data on patient characteristics, hospital course, laboratory values, vital signs, surgical procedures, and medication administration were collected on all study enrolled patients during the course of their hospitalization. The validated confusion assessment method for the ICU (CAM-ICU) screening tool was used to diagnose delirium in our study population. [25] The CAM-ICU tool is administered by Mayo Clinic ICU nursing staff every 8 h or when a mental status change is noted. Positive and negative CAM-ICU results are documented in the electronic health record as events occurring during the following 5-h windows: 0:00–4:00; 4:00–8:00; 8:00–12:00; 12:00–16:00; 16:00–20:00; and 20:00–24:00. The prerequisite to inclusion in the delirium group was positive screening test on two consecutive administrations of the CAM-ICU.

For the purpose of our study, we defined delirium in accordance with the American Psychiatric Association's Diagnostic and Statistical of Mental Disorders, 5th edition.[26] That definition states that delirium is a disturbance of attention, cognition, and awareness that develops over a short period of time, represents an acute change from baseline attention and awareness, and tends to fluctuate in severity throughout the course of a day. Such disturbance must not better explained by another preexisting, evolving, or established neurocognitive disorder and does not occur in the context of a severely reduced level of arousal, such as coma. Finally, the definition requires evidence from the patient's history, physical examination, or laboratory results that the disturbance is caused by a medical condition, substance intoxication or withdrawal, or medication side effect. Furthermore, we utilized the Third International Consensus Definitions (Sepsis-3) to define both sepsis and septic shock.[27]

In those patients that underwent surgical procedures, inhaled anesthetic exposure, medication administration, and surgery characteristics data were also collected. Vital signs and laboratories were collected during the first ICU day; worse values were abstracted. The acute physiology score (APS), acute physiology and chronic health evaluation score (APACHE IV), and predicted hospital mortality rates based on these scores were calculated using an online APACHE IV calculator.[28] Daily sequential organ failure assessment score[29] was documented daily from day 1 to day 7. Unobtainable missing data would have resulted in list-wise deletion; however, this strategy was not required in our dataset.

Statistics

All data are summarized as median (interquartile range) or percentages. Unpaired Student's t-tests will be used to compare continuous variables with normal distribution and the Wilcoxon rank test for skewed distribution. For comparison of categorical variables, Chi-square tests will be used if the number of elements in each cell was ≥5; Fisher's exact test will be used otherwise. For the assessment of independent predictors of delirium, we created a multiple logistic regression model by entering covariates that showed significant differences (P < 0.1) between those that develop or not delirium. The model was refined with backward stepwise regression, considering colinearity and interaction terms. When appropriate, the odds ratio (OR) and 95% confidence intervals (CI) were calculated. Model discrimination was assessed using receiving operator curves. Model fit (calibration) was assessed using the Hosmer-Lemeshow goodness-of-fit test. A P ≤ 0.05 was considered statistically significant. All data analyses were performed using JMP Statistical Software (Version 14.1.0; SAS Institute, Cary, NC, USA).

RESULTS

A total of 162 patients were included in the study; 50 of whom developed delirium during their hospitalization. Description of the study population, data on the hospital course of each group, as well as laboratory values and vital signs are provided in Tables 1-3, respectively. The study cohort includes 39 (24.1%) COVID-19-positive patients. Univariate analysis suggested that COVID-19 was more commonly observed in the delirium population. Both APS and APACHE IV were higher in the group that developed delirium. Sepsis and septic shock were more commonly associated with the delirium group in the univariate analysis. Furthermore, continuous renal replacement therapy (CRRT), extracorporeal membrane oxygenation (ECMO), immunosuppression, tachycardia, tachypnea, mechanical ventilation, leukocytosis, and acidemia were also more commonly identified in the population that developed delirium. The delirium group also experienced increased ICU and hospital lengths of stay (LOS).

Table 1

Patient characteristics

Variable	No delirium (n=112), n (%)	Delirium (n=50), n (%)	P	Total cohort
Age (years), median (IQR)	66.5 (55.2-75)	55.5 (47-70.2)	0.005*	64.5 (51-75)
Female gender, n (%)	37 (33)	18 (36)	0.72	55 (34)
APS score, median (IQR)	38 (26.3-48.8)	62.5 (41.5-80.3)	<0.001*	44.5 (30-59.3)
APACHE IV, median (IQR)	53 (37.3-66)	80.5 (53.8-96.8)	<0.001*	56 (39-75.3)
CVA (ischemic stroke)	10 (8.93)	7 (14)	0.41	17 (10.49)
Hemorrhagic stroke	1 (0.89)	2 (4)	0.23	3 (1.85)
TBI	2 (1.79)	1 (2)	1.0	3 (1.85)
Trauma	1 (0.89)	0	1.0	1 (0.62)
Dementia	2 (1.79)	1 (2)	1.0	3 (1.85)
BMI, median (IQR)	28.2 (24.5-33.5)	29.3 (25.4-34.3)	0.50	28.9 (24.9-33.6)
Code status (full code)	110 (98.2)	50 (100)	1.0	160 (98.8)
DM	35 (31.25)	20 (40)	0.29	55 (34)
HTN	49 (43.75)	21 (42)	0.87	70 (43.2)
CAD	29 (25.9)	11 (22)	0.69	40 (24.7)
Stroke	10 (8.9)	6 (12)	0.57	16 (9.9)
Cancer	21 (18.75)	9 (18)	1.0	30 (18.5)
CKD	16 (14.3)	4 (8)	0.31	20 (12.4)
ESRD	4 (3.6)	0	0.31	4 (2.5)
Cirrhosis	8 (7.1)	1 (2)	0.28	9 (5.6)
Hepatic failure	1 (0.9)	0	1.0	1 (0.6)
Metastatic carcinoma	4 (3.6)	4 (8)	0.25	8 (4.9)
Lymphoma	2 (1.8)	0	1.0	2 (1.2)
Leukemia/myeloma	3 (2.7)	2 (4)	0.64	5 (3.1)
Immunosuppression	9 (8)	10 (20)	0.04*	19 (11.7)

*P≤0.05. APACHE IV: Acute physiology and chronic health evaluation score, APS: Acute physiology score, BMI: Body mass index, CAD: Coronary artery disease, CKD: Chronic kidney disease, CVA: Cerebrovascular accident, DM: Diabetes mellitus, ESRD: End stage renal disease, HTN: Hypertension, IQR: Interquartile range, TBI: Traumatic brain injury

Table 3

Labs and vitals

Variable	No delirium (n=112), n (%)	Delirium (n=50), n (%)	P	Total cohort
Positive cultures	23 (20.5)	14 (28)	0.32	37 (22.8)
Initial lactate (mmol/L), median (IQR)	1.5 (1.1-2.6)	1.8 (1.2-3.8)	0.24	1.5 (1.1-3)
Highest lactate (mmol/L), median (IQR)	1.8 (1.1-3.6)	2.2 (1.4-4.5)	0.16	1.9 (1.2-3.8)
Lactate clearance (mmol/L), median (IQR)	1.6 (1.3-2.9)	2.3 (1.4-3.2)	0.38	1.7 (1.3-3.2)
Ionized calcium (mg/dL), median (IQR)	4.5 (4.3-4.8)	4.4 (4.1-4.6)	0.20	4.5 (4.2-4.7)
Calcium (mg/dL), median (IQR)	8.4 (7.8-9)	8.2 (7.5-8.5)	0.05*	8.2 (7.8-8.8)
Temperature (°C), median (IQR)	37.4 (36.7-38.5)	37.4 (36.8-38.2)	0.78	37.4 (36.7-38.3)
MAP (mmHg), median (IQR)	62 (56.3-73)	64 (58-70.3)	0.51	62 (57.8-71.3)
HR, median (IQR)	98.5 (88-113.8)	105.5 (92.8-122.3)	0.03*	101 (90-116.3)
RR, median (IQR)	24 (20-29)	27 (23.8-32)	0.04*	25 (20-29.3)
MV	45 (40.2)	33 (66)	0.004*	78 (48.2)
FiO2%, median (IQR)	50 (28-60)	60 (50-100)	<0.001*	50 (31.5-60)
PaO2 (mmHg), median (IQR)	90 (80-106.8)	89.5 (73.5-105)	0.59	90 (78-105.3)
PaCO2 (mmHg), median (IQR)	37.5 (32-40)	36.5 (31-43.3)	0.43	37 (32-40)
pH art, median (IQR)	7.4 (7.37-7.43)	7.38 (7.29-7.43)	0.04*	7.4 (7.35-7.43)
Na+ (mmol/L), median (IQR)	138 (135-141)	136 (132-140)	0.05*	138 (134-141)
Urine output (mL/24 h), median (IQR)	1571.5 (1028.8-2132.5)	1662.5 (923.8-2376.3)	0.88	1592 (987.5-2250.8)
Creatinine (mg/dL), median (IQR)	1.1 (0.82-1.62)	1.34 (0.83-2.02)	0.14	1.1 (0.83-1.67)
BUN (mg/dL), median (IQR)	17 (12.3-26.1)	24.5 (14-40.3)	0.02*	18.4 (13-29.9)
Glucose (mg/dL), median (IQR)	142.5 (112-176)	166 (138.8-154.8)	0.001*	149.5 (114.8-185)
Albumin (g/dL), median (IQR)	3.9 (3.2-4)	3.3 (2.8-3.8)	0.002*	3.6 (3-4)
Bilirubin (mg/dL), median (IQR)	0.8 (0.43-1)	0.6 (0.3-1.03)	0.17	0.7 (0.4-1)
Hematocrit percentage, median (IQR)	33.9 (27.9-38.1)	32.1 (25.8-38.9)	0.54	33.1 (27.5-38.2)
WBC, median (IQR)	9.5 (6.8-13.5)	12.9 (9.2-23.9)	0.003*	10.3 (7.1-15.1)

*P≤0.05. BMI: Body mass index, BUN: Blood urea nitrogen, FiO2%: Fraction of inspired oxygen, HR: Heart rate, IQR: Interquartile range, MAP: Mean arterial pressure, Na+: Ionized sodium, PaCO2: Partial pressure of carbon dioxide, PaO2: Partial pressure of oxygen, pH art: Arterial pH, RR: Respiration rate, WBC: White blood cell

Information related to surgical risk factors associated with delirium in our patient population is described in Table 4. Although patients who did not experience delirium were more likely to have had recent surgery and inhaled anesthetic exposure, no significant difference was seen between the nondelirium and delirium groups when considering blood loss from all surgical procedures, total surgical time, and total duration of inhaled anesthetic exposure.

Table 4

Surgical risk factors

Variable	No delirium	Delirium	P	Total cohort
Recent surgery, n (%)	51 (45.5)	13 (26)	0.02*	64 (39.5)
Total blood loss from all surgical procedures (mL), median (IQR)	100 (75-400)	125 (100-400)	0.52	100 (100-395)
Total surgical time from all surgical procedures (min), median (IQR)	302 (241-383)	284 (223-501.5)	0.62	300.5 (238.3-408.3)
Recent inhaled anesthetic, n (%)	54 (48.2)	14 (28)	0.02*	68 (42)
Total duration of inhaled anesthesia from all surgical procedures (min), median (IQR)	293 (197-382.3)	255 (214-522.5)	0.49	287 (207-410)

*P≤0.05. IQR: Interquartile range

Potentially deliriogenic medications administered to this patient population during the course of their hospitalization are provided in Tables 5 and 6. Those patients that received diazepam and midazolam were more likely to develop delirium; this relationship held with predelirium versus total dose, given in lorazepam equivalents, of both diazepam and midazolam. Propofol dose as well as dexmedetomidine use and dose were also associated with the development of delirium. Total opioid dose, reported in fentanyl equivalents, was associated with the development of delirium; this was predominantly driven by fentanyl and hydromorphone administration. Cefepime, vancomycin, azithromycin, and piperacillin use were also associated with the development of delirium; however, no significance was seen when considering the dose of these medications. Neither use nor total dose of corticosteroids or medications with significant anticholinergic properties was associated with the development of delirium in our study population.

Table 5

Benzodiazepines, sedatives, opioids

Variable	No delirium	Delirium	P	Total cohort
Benzodiazepines (in lorazepam equivalents)
Diazepam, n (%)	7 (6.3)	13 (26)	<0.001*	20 (12.4)
Diazepam dose (mg), median (IQR)	3 (1.7-13.7)	24.1 (12.1-36.3)	0.04*	16.7 (3.5-29.2)
Midazolam, n (%)	67 (59.8)	41 (82)	0.006*	108 (66.7)
Midazolam dose (mg), median (IQR)	3.5 (2-5.7)	101 (5.5-966.7)	<0.001*	5 (2.2-89.9)
Lorazepam, n (%)	24 (21.4)	6 (12)	0.19	30 (18.5)
Lorazepam dose (mg), median (IQR)	1.5 (0.5-4)	1 (0.8-21.3)	0.98	1.5 (0.5-3.5)
Alprazolam, n (%)	7 (6.3)	2 (4)	0.72	9 (5.6)
Alprazolam dose (mg), median (IQR)	3.6 (1.0-5)	2 (2-2)	1.0	2.8 (1-4.8)
Sedatives
Etomidate, n (%)	29 (25.9)	19 (38)	0.14	48 (29.6)
Etomidate dose (mg), median (IQR)	24 (20-30)	20 (20-22.8)	0.08	20 (20-30)
Methocarbamol, n (%)	3 (2.7)	3 (6)	0.37	6 (3.7)
Methocarbamol dose (mg), median (IQR)	6875 (3187.6-7000)	1500 (500-10,000)	0.59	6750 (1500-7000)
Propofol, n (%)	50 (44.6)	28 (56)	0.23	78 (48.1)
Propofol dose (mg), median (IQR)	260.5 (120-1071.5)	1980 (150.3-4136)	0.05*	486 (138.4-3331)
Dexmedetomidine, n (%)	43 (38.4)	29 (58)	0.03*	72 (44.4)
Dexmedetomidine dose (mcg), median (IQR)	397 (147.2-820)	2484.8 (970.8-7992)	<0.001*	622 (243.3-2540)
Ketamine, n (%)	9 (8)	13 (26)	0.005*	22 (13.6)
Ketamine dose (mg), median (IQR)	287 (20-2666.8)	2553.5 (50-4422.4)	0.28	1184.7 (30-3846.3)
Opioids (in fentanyl equivalents)
Fentanyl, n (%)	82 (73.2)	43 (86)	0.10	125 (77.2)
Fentanyl dose (mcg), median (IQR)	1175 (200-1962.5)	2400 (450-10,093)	0.002*	1362.5 (256.3-2568.8)
Hydromorphone, n (%)	39 (34.8)	22 (44)	0.29	61 (37.7)
Hydromorphone dose (mcg), median (IQR)	120 (40-800)	2553.6 (40-85,640.5)	0.03*	204 (40-2693.6)
Morphine, n (%)	14 (12.5)	9 (18)	0.46	23 (14.2)
Morphine dose (mcg), median (IQR)	95 (40-270)	100 (50-1266.3)	0.78	100 (40-480)
Oxycodone, n (%)	50 (44.6)	8 (16)	<0.001*	58 (35.8)
Oxycodone dose (mcg), median (IQR)	975 (375-1668.8)	4500 (225-1312.5)	0.41	975 (375-1593.8)
Tramadol, n (%)	9 (8)	3 (6)	0.76	12 (7.4)
Tramadol dose (mcg), median (IQR)	150 (75-425)	150 (100-200)	0.64	150 (62.5-375)
Total fentanyl equivalents (mcg), median (IQR)	1962.5 (427.3-3715)	5478.8 (750.6-32,795.8)	0.002*	2368 (531.3-5153.8)

*P≤0.05. IQR: Interquartile range

Table 6

Antibiotics, steroids, and drugs with anticholinergic properties

Variable	No delirium	Delirium	P	Total cohort
Antibiotics
Cefazolin, n (%)	51 (45.5)	13 (26)	0.02*	64 (39.5)
Cefazolin dose (g), median (IQR)	12 (6–12)	8 (3–12)	0.12	12 (6–12)
Cefepime, n (%)	18 (16.1)	16 (32)	0.04*	34 (21)
Cefepime dose (g), median (IQR)	12 (4.8–17.3)	12 (4–26)	0.59	12 (4.5–19)
Vancomycin, n (%)	46 (41.1)	37 (74)	<0.001*	83 (51.2)
Vancomycin dose (g), median (IQR)	2.8 (1.8–5.1)	3 (1.8–5)	0.82	3 (1.8–5)
Acyclovir, n (%)	4 (3.6)	4 (8)	0.25	8 (4.9)
Acyclovir dose (g), median (IQR)	4 (1.6–22.1)	3.6 (2.3–7.8)	1.0	3.6 (2.3–8)
Azithromycin, n (%)	22 (19.6)	21 (42)	0.004*	43 (26.5)
Azithromycin dose (g), median (IQR)	1.5 (0.5–2.5)	1.5 (0.5–2.5)	0.80	1.5 (0.5–2.5)
Piperacillin, n (%)	37 (33)	34 (68)	<0.001*	71 (43.8)
Piperacillin dose (g), median (IQR)	37.1 (27–93.4)	34.9 (12.7–59.9)	0.32	37.1 (20.3–87.8)
Ceftriaxone, n (%)	26 (23.2)	8 (16)	0.40	34 (21)
Ceftriaxone dose (g), median (IQR)	2 (1–6)	5.5 (3.3–6.8)	0.19	4 (1–6)
Ertapenem, n (%)	5 (4.5)	1 (2)	0.67	6 (3.7)
Ertapenem dose (g), median (IQR)	5 (2.5–7.5)	6	1.0	5.5 (2.8–7.3)
Amoxicillin, n (%)	8 (7.1)	1 (2)	0.28	9 (5.6)
Amoxicillin dose (g), median (IQR)	3.1 (1.6–5)	2.6	1.0	2.6 (1.6–4.8)
Ciprofloxacin, n (%)	3 (2.7)	2 (4)	0.64	5 (3.1)
Ciprofloxacin dose (g), median (IQR)	1 (0.8–3)	6 (5.2–6.7)	0.15	3 (0.9–6)
Sulfamethoxazole, n (%)	5 (4.5)	1 (2)	0.67	6 (3.7)
Sulfamethoxazole dose (g), median (IQR)	2.4 (1.6–3.4)	0.8	0.20	2.4 (1.6–3.2)
Trimethoprim, n (%)	6 (5.4)	1 (2)	0.44	7 (4.3)
Trimethoprim dose (g), median (IQR)	0.5 (0.3–0.7)	0.2	0.20	0.5 (0.3–0.6)
Caspofungin, n (%)	7 (6.3)	6 (12)	0.22	13 (8)
Caspofungin dose (g), median (IQR)	0.3 (0.1–0.5)	0.4 (0.1–0.5)	0.87	0.3 (0.1–0.4)
Fluconazole, n (%)	17 (15.2)	8 (16)	1.0	25 (15.4)
Fluconazole dose (g), median (IQR)	1.2 (0.8–2.1)	1.2 (0.5–3.8)	0.77	1.2 (0.8–2.3)
Ampicillin, n (%)	3 (2.7)	2 (4)	0.64	5 (3.1)
Ampicillin dose (g), median (IQR)	63 (18–72)	16.5 (12–21)	0.39	21 (15–67.5)
Doxycycline, n (%)	5 (4.5)	3 (6)	0.70	8 (4.9)
Doxycycline dose (g), median (IQR)	0.8 (0.2–1.6)	0.7 (0.2–1.4)	1.0	0.8 (0.2–1.3)
Metronidazole, n (%)	13 (11.6)	5 (10)	1.0	18 (11.1)
Metronidazole dose (g), median (IQR)	1.5 (1–4)	3 (1.8–9.5)	0.23	2.5 (1–4.1)
Levofloxacin, n (%)	3 (2.7)	2 (4)	0.64	5 (3.1)
Levofloxacin dose (g), median (IQR)	0.8 (0.5–0.8)	0.6 (0.5–0.8)	1.0	0.8 (0.5–0.8)
Meropenem, n (%)	13 (11.6)	9 (18)	0.32	22 (13.6)
Meropenem dose (g), median (IQR)	7 (2.5–17.5)	14 (3–35.5)	0.42	7 (3–21.3)
Clindamycin, n (%)	3 (2.7)	1 (2)	1.0	4 (2.5)
Clindamycin dose (g), median (IQR)	3.6 (2.7–5.1)	1.8	0.37	3.2 (2–4.7)
Tobramycin, n (%)	2 (1.8)	2 (4)	0.59	4 (2.5)
Tobramycin dose (g), median (IQR)	2.3 (0.1–4.6)	0.6 (0.3–0.9)	1.0	0.6 (0.1–3.7)
Valacyclovir, n (%)	1 (0.9)	1 (2)	0.52	2 (1.2)
Valacyclovir dose (g), median (IQR)	3	4	1.0	3.5 (3–4)
Daptomycin, n (%)	2 (1.8)	1 (2)	1.0	3 (1.9)
Daptomycin dose (g), median (IQR)	1.8 (0.9–2.8)	4.6	0.54	0.9 (0.5–2.8)
Ceftolozane-tazobactam, n (%)	1 (0.9)	1 (2)	0.52	2 (1.2)
Ceftolozane-tazobactam dose (g), median (IQR)	111	57	1.0	84 (57–111)
Hydroxychloroquine, n (%)	5 (4.5)	3 (6)	0.70	8 (4.9)
Hydroxychloroquine dose (g), median (IQR)	2.2 (1.3–2.4)	0.8 (0.6–2.4)	0.44	2 (0.8–2.4)
Remdesivir, n (%)	8 (7.1)	6 (12)	0.37	14 (8.6)
Remdesivir dose (g), median (IQR)	0.6 (0.5–0.6)	0.6 (0.5–0.7)	0.94	0.6 (0.5–0.6)
Steroids
Hydrocortisone, n (%)	37 (33)	23 (46)	0.16	60 (37)
Hydrocortisone dose (mg), median (IQR)	400 (275–600)	425 (150–1000)	0.91	400 (250–650)
Methylprednisolone, n (%)	7 (6.3)	5 (10)	0.52	12 (7.4)
Methylprednisolone dose (mg), median (IQR)	250 (100–395)	75 (40–616)	0.52	187.5 (66.8–548.8)
Dexamethasone, n (%)	9 (8)	3 (6)	0.76	12 (7.4)
Dexamethasone dose (mg), median (IQR)	60 (4–122)	14 (10–18)	0.85	16 (4.5–60)
Prednisone, n (%)	6 (5.4)	4 (8)	0.50	10 (6.2)
Prednisone dose (mg), median (IQR)	65 (23.8–145)	25 (20–202.5)	0.59	35 (20–145)
Drugs with anticholinergic properties
Famotidine, n (%)	9 (8)	4 (8)	1.0	13 (8)
Famotidine dose (mg), median (IQR)	60 (20–140)	20 (20–110)	0.33	40 (20–140)
Diphenhydramine, n (%)	14 (12.5)	4 (8)	0.59	18 (11.1)
Diphenhydramine dose (mg), median (IQR)	75 (25–112.5)	137.5 (53.1–231.3)	0.31	75 (34.4–156.3)
Prochlorperazine, n (%)	7 (6.3)	2 (4)	0.72	9 (5.6)
Prochlorperazine dose (mg), median (IQR)	10 (5–20)	10	0.88	10 (7.5–17.5)
Promethazine, n (%)	1 (0.9)	2 (4)	0.23	3 (1.9)
Promethazine dose (mg), median (IQR)	12.5	78.1 (6.3–150)	1.0	12.5 (6.3–150)

*P≤0.05. IQR: Interquartile range

The multivariate analysis for the evaluation of independent association of opioids, benzodiazepines, severity of disease, age, and presence of COVID-19 with the development of delirium is presented in Table 7. Only opioids, benzodiazepines, and estimated mortality by APACHE IV were independently associated with the development of delirium. Neither COVID-19 nor age was associated with the development of delirium after adjustment for important covariates.

Table 7

Multivariate analysis, receiving operator curves model 0.97

Variable	OR	95% CI	P
Opioids	24	16-27	<0.001*
Benzodiazepines	23	16-63	0.001*
Estimated mortality	1.04	1.01-1.07	0.002*
Age	0.97	0.92-1.03	0.3
COVID-19	1.44	0.13-10.6	0.7

*P≤0.05. CI: Confidence interval, OR: Odds ratio, COVID-19: Coronavirus disease-19

DISCUSSION

In this prospective cohort of critically ill patients admitted to the ICU, we identified that delirium was prevalent in COVID-19 patients, with 20 (51.3%) of those in our study developing this syndrome. However, after adjustment for important confounders, our data showed that SARS-CoV-2 infection was not an independent risk factor for the development of delirium. The association of critical illness, indicated here through APS and APACHE IV, with the development of delirium was established.[791113] Many of the aforementioned variables that demonstrated a significant association with delirium are common to critically ill patients treated in the ICU. Such factors include the following: use of CRRT, ECMO, and mechanical ventilation; immunosuppression; vital signs such as tachycardia and tachypnea; and laboratory abnormalities including leukocytosis and acidemia. Sepsis and septic shock, like severe COVID-19, result in significant systemic inflammation; it is the release of proinflammatory cytokines associated with critical illness that is thought to be the proximate cause of delirium.[30313233] Based on our findings, COVID-19 seems to be associated with the development of delirium in the critically ill due to other factors associated with illness severity in this condition and not because of some unique property intrinsic to SARS-CoV-2. It remains possible that in other phenotypes of COVID-19, such as less severe illness or CNS predominant disease, direct CNS inflammation may lead to delirium, as well as chronic encephalopathy or neuropsychiatric sequela.

The exact mechanism by which COVID-19 contributes to the development of delirium remains to be elucidated.[2] The effects of SARS-CoV-2 on the brain are likely multifactorial, with insult occurring both directly and indirectly.[1234910182234353637] Direct CNS invasion is theoretically possible through either retrograde axonal transport or hematogenous spread.[210133435] Considering the common finding of anosmia,[21923] at the outset of the pandemic, there were concerns that SARS-CoV-2 could result in encephalitis by infecting olfactory neurons.[29] However, subsequent studies have shown that these symptoms likely result from an infection of supporting cells in the epithelium.[23]

Alternatively, direct neuroinvasion may occur due to the presence of angiotensin-converting enzyme 2, the receptor to which the virus spike protein binds on neurons, astrocytes, and oligodendrocytes.[2379111314] Those cells found in the circumventricular organs are thought to be particularly susceptible to infection due to vascularization and permeability of the local blood–brain barrier.[291219] However, despite case reports of encephalitis,[1935] the paucity of patients in which SARS-CoV-2 is detected in the cerebrospinal fluid by PCR suggests other factors play a more significant role in the development of delirium than direct neuroinvasion.[121923] Furthermore, COVID-19 is associated with systemic coagulopathy and vasculopathy, resulting in macro-and microvascular pathology, both of which may contribute to the neurocognitive findings seen in these patients.[12323]

Severe COVID-19 results in significant systemic inflammation.[14921] The release of cytokines and chemokines increases blood–brain barrier permeability and activates resident neuroglial cells; both events are associated with the development of delirium.[129192324] Although the pathogenesis of severe COVID-19 has been described as being contributed to by the occurrence of a cytokine storm,[791623] this phenomenon is not unique to SARS-CoV-2 infection; similar cytokine levels are seen in ARDS and sepsis secondary to bacterial infection.[118] It is hypothesized that the preeminent cause of delirium in COVID-19 patients results from this mechanism.[6710192123]

Opioids, benzodiazepines, and certain sedatives are known to be deliriogenic;[67153138394041] these medications are commonly used in the ICU. Our study shows a relationship between both diazepam and midazolam usage and dose with the development of delirium. Propofol showed a dose, but not use, related association with delirium. We suspect that this results from the dual utilization of this medication to induce anesthesia and for prolonged ICU sedation; during which significantly greater doses are administered. Both fentanyl and hydromorphone dose were correlated with the development of delirium, as was total opioid administration. Both opioids and benzodiazepines withstood multivariate analysis regarding their contributions to the development of delirium. Interestingly, anticholinergic medications, which have been associated with delirium,[1542] did not increase the risk of delirium in this population. A possible explanation may be related to cognitive reserve (i.e., baseline cognitive resilience). In this cohort, younger and more severely ill patients developed delirium; anticholinergic medications more commonly increase the risk of delirium in older patients who are more likely to have underlying cognitive impairment or a lower cognitive reserve.[42] In addition, age has been identified as an independent risk factor for the development of delirium.[15] Initial analysis showed that, in our patient population, the age of patients that did not develop delirium was greater compared to those who did. However, after adjusting for LOS (longer LOS resulted in a higher risk of developing delirium; surgical patients had a shorter LOS and were significantly younger), the age difference was no longer noted between the two groups.

Our study had several limitations including its modest size, performance at a single center, and imbalance between the size of the delirium and nondelirium groups. An additional limitation arose from the inclusion of surgical patients in our study population. Surgery is a well-established risk factor for the development of delirium; this is also an inflammatory cytokine-driven process.[4344] However, those patients in our study who recently had surgery were less likely to experience delirium. This unexpected finding is explained by the elective nature of many of the surgical procedures performed on older patients that required only shorter duration ICU and hospital courses. This had the effect of limiting many of the predisposing factors that otherwise contribute to the development of this syndrome, in our surgical patients. We acknowledge this limitation of our study and suggest thatfuture studies confirm our results in a patient population consisting of exclusively nonsurgical patients. In addition, a multicenter prospective study involving a large patient population is necessary to confirm and further characterize the association between COVID-19 and delirium.

CONCLUSION

Although delirium is prevalent in COVID-19 patients, our data suggest that SARS-CoV-2 infection in critically ill patients is not an independent risk factor for its development. It is critical illness itself, denoted by APACHE IV, that accounts for the high rates of delirium in COVID-19 patients. Investigation of COVID-19-related cytokine release, which is associated with the development of delirium in septic shock and other infectious etiologies of critical illness,[30313233] may further elucidate the mechanism by which SARS-CoV-2 causes delirium in these patients.

Research quality and ethics statement

This study was approved by the IRB at the Mayo Clinic (Approval # 18-005104; Approval date January 26, 2019). The authors followed the applicable EQUATOR Network (http://www.equator-network.org) guidelines, specifically the STROBE Guidelines, during the conduct of this research project.

Financial support and sponsorship

None.

Conflicts of interest

There are no conflicts of interest.

Table 2

Hospital course

Variable	No delirium (n=112), n (%)	Delirium (n=50), n (%)	P	Total cohort
ICU LOS, median (IQR)	2.1 (1.2-3.7)	13.6 (3.4-30.1)	<0.001*	2.97 (1.83-8.26)
Hospital LOS, median (IQR)	7.1 (4.9-11.5)	23 (12.8-38.2)	<0.001*	9.7 (5.5-20.3)
ICU mortality	8 (7.14)	7 (14)	0.24	15 (9.26)
Hospital mortality	8 (7.14)	8 (16)	0.09	16 (9.9)
Septic shock	29 (25.89)	22 (44)	0.028*	51 (31.5)
Sepsis or septic shock	49 (43.8)	34 (68)	0.006*	83 (51.2)
COVID-19	19 (17)	20 (40)	0.003*	39 (24.1)
28-day mortality	9 (8.04)	9 (18)	0.10	18 (11.11)
GCS, median (IQR)	15 (12-15)	11 (6-15)	<0.001*	14.5 (10-15)
Pre-ICU days	0 (0-2)	1 (0-2)	0.39	0 (0-2)
Emergency surgery	4 (3.6)	2 (4)	1.0	6 (3.7)
Thrombolysis	14 (12.5)	1 (2)	0.04*	15 (9.3)
CRRT	7 (6.3)	14 (28)	0.001*	21 (13)
ECMO	1 (0.9)	10 (20)	<0.001*	11 (6.8)
LVAD	1 (0.9)	1 (2)	0.52	2 (1.2)
RVAD	1 (0.9)	1 (2)	0.52	2 (1.2)

*P≤0.05. COVID-19: Coronavirus disease-19, CRRT: Continuous renal replacement therapy, ECMO: Extracorporeal membrane oxygenation, GCS: Glasgow Coma Scale, ICU: Intensive care unit, IQR: Interquartile range, LOS: Length of stay (days), LVAD: Left ventricular assist device, RVAD: Right ventricular assist device