Journal of Critical Care 29 (2014) 517–522

Contents lists available at ScienceDirect

Journal of Critical Care

j ourna l homepage: www. jcc journa l .org

Effects of etomidate on vasopressor use in patients with sepsis or

severe sepsis: A propensity-matched analysis☆,☆☆

Nerissa J. Alday, PharmD b, G. Morgan Jones, PharmD, BCPS c, Lauren A. Kimmons, PharmD, BCPS c,Gary S. Phillips, MAS d, Jennifer W. McCallister, MD e,f, Bruce A. Doepker, PharmD, BCPS a,⁎a Department of Pharmacy, Wexner Medical Center at Ohio State University, Columbus, OH 43210-1228b Department of Pharmacy, Nationwide Children's Hospital, Columbus, OH 43205c Methodist University Hospital, University of Tennessee College of Pharmacy, Memphis, TN 38104d Center for Biostatistics, Ohio State University, Columbus, OH 43221e Department of Internal Medicine, Wexner Medical Center at Ohio State University, Columbus, OH 43221f Pulmonary, Allergy, Critical Care and Sleep Medicine, Davis Heart and Lung Research Institute, Columbus, OH 43210-1267

a b s t r a c ta r t i c l e i n f o

☆ Financial disclosure: The authors have no actual or prelation to the conduct of this study.☆☆ Funding: None.⁎ Corresponding author. Department of Pharmacy, W

State University, Columbus, OH; 410 W. 10th Ave, Room43210-1228. Fax: +1 614 293 3165.

E-mail address: Bruce.Doepker@osumc.edu (B.A. Do

http://dx.doi.org/10.1016/j.jcrc.2014.02.0020883-9441/© 2014 Elsevier Inc. All rights reserved.

Keywords:

EtomidateSepsisAdrenal insufficiencyHypotensionVasopressor

Purpose: The safety of single-bolus etomidate to facilitate intubation in septic patients is controversial dueto its potential to suppress adrenal steroidogenesis. The purpose of this study was to evaluate the effectsof etomidate on the development of shock when used as an induction agent to facilitate intubation inseptic patients.Methods: A multicenter, retrospective, propensity-matched cohort study comparing patients with sepsis or

severe sepsis who either received etomidate or did not receive etomidate for intubation was conducted.The primary outcome was the difference in the need for vasopressor support within 72 hours afterintubation. Secondary outcomes included the use of multiple vasopressors, intensive care unit length ofstay, and in-hospital mortality.Results: A total of 411 patients were analyzed. Eighty-three patients were matched by propensity score.There was no difference in the matched cohort in regards to vasopressor use within 72 hours of intubation(odds ratio, 0.95; 95% confidence interval, 0.52-1.76; P = .88). Furthermore, there were no significantdifferences observed with regard to secondary outcomes, including in-hospital mortality (P = .76).Conclusions: The use of etomidate for intubation in septic patients did not increase vasopressorrequirements within 72 hours after intubation.

© 2014 Elsevier Inc. All rights reserved.

1. Background

Etomidate is an imidazole sedative-hypnotic agent used tofacilitate intubation. It has a rapid, predictable onset of action andrecovery time compared with other induction agents. Historically,etomidate has been a favorable agent for use in patients withhemodynamic compromise due to its low rate of cardiovascular sideeffects and minimal risk of respiratory depression [1-3]. However,recent literature has created controversy regarding the use ofetomidate in patients with hemodynamic compromise secondary tosepsis because of its potential to induce adrenal insufficiency andincrease the risk of hypotension.

otential conflicts of interest in

exner Medical Center at Ohio368 Doan Hall, Columbus, OH

epker).

Etomidate is known to suppress adrenal steroidogenesis for up to72 hours after a single dose through the inhibition of 11-β-hydroxylase, a key enzyme involved in the conversion of deoxycorti-sol to cortisol [2-7].

In septic patients, this inhibitionmay blunt cortisol production that isstimulated by activation of the hypothalamic-pituitary-adrenal axis.Because septic patients attempt to maintain vascular homeostasis bymounting a compensatory adrenal stress response, those receivingetomidatemaybe at an increased risk for refractoryhypotension. Severallarge studies of septic patients have raised concern over the associationbetween etomidate-related adrenal insufficiency and increased mortal-ity; however, these studies were not adequately designed to evaluatethis outcome [2,8-12]. More recent retrospective studies have failed toreproduce these earlier findings [7,13-16]. There is also a lack of qualityevidence evaluating outcomes such as vasopressor use and hypotensionassociated with etomidate-related adrenal insufficiency in patients withsepsis and severe sepsis. Published studies that have reportedhemodynamic outcomes associated with single-bolus etomidate havebeen primarily observational in nature and included small patientpopulations with severe sepsis and septic shock [7,13,14,16,17].

518 N.J. Alday et al. / Journal of Critical Care 29 (2014) 517–522

The impact of single-bolus dose etomidate on outcomes such asvasopressor use, hypotension, and mortality in patients with sepsis orsevere sepsis remains to be fully characterized. This study serves asthe first propensity-matched study designed to evaluate the hemo-dynamic consequences of etomidate use before progression to septicshock. The primary objective was to further define the associationbetween etomidate use and postintubation vasopressor requirementsamong patients with sepsis or severe sepsis.

2. Methods

2.1. Study design

A retrospective, multicenter propensity-matched cohort study ofpatients with sepsis or severe sepsis was performed at the Ohio StateUniversity Wexner Medical Center, a 1200-bed academic tertiary careinstitution in Columbus, OH, and Methodist University Hospital, a660-bed community academic medical center in Memphis, TN.Patients were eligible for inclusion if they were between the ages of18 and 89 years, had a diagnosis of suspected or confirmed sepsis orsevere sepsis, and were intubated at either institution betweenJanuary 1, 2007, and October 31, 2012. Exclusion criteria includedincarceration, pregnancy, history of chronic adrenal insufficiency,vasopressor use within 24 hours before intubation, corticosteroid usewithin 30 days before intubation, history of chronic immunosuppres-sion (including HIV/AIDS, chemotherapy, radiation therapy, ortransplant immunosuppressive therapy within 30 days beforeintubation), hypersensitivity to etomidate, or an advanced directiveto withhold or withdraw life-sustaining treatment before intuba-tion. This study was approved by institutional review boards atWexner Medical Center and the University of Tennessee HealthSciences Center.

2.2. Data collection and measurement

Patients were identified using an electronic database at bothmedical centers to select for patients with International Classificationof Diseases, Ninth Revision (ICD-9), codes for diagnoses of sepsis orsevere sepsis and mechanical ventilation. Chart review usingelectronic medical records and paper charts was used to screenpatients for inclusion and to extract information for analysis. Thefollowing baseline variables were recorded within 24 hours beforeintubation: demographics, presence of sepsis or severe sepsis criteria,source of infection, community-acquired infection or health care–associated infection, Acute Physiology and Chronic Health Evaluation(APACHE) II score, Sequential Organ Failure Assessment (SOFA) score,and Glasgow Coma Scale (GCS) score. Vital signs included baselinemean arterial pressure (MAP) and systolic blood pressure (SBP)within 2 hours before intubation as well as the lowest reported MAPand SBP within 0 to 12, 12 to 24, 24 to 48, and 48 to 72 hourspostintubation. Additional data included induction agent(s) and dose(s) administered, vasopressor (epinephrine, norepinephrine, phenyl-ephrine, dopamine, or vasopressin) use and maximum infusion rateswithin 72 hours postintubation, duration of mechanical ventilation,hospital and intensive care unit (ICU) length of stay (LOS), all-causemortality, serum cortisol levels, and postintubation corticosteroid use.The decision to intubate, choice of agents for induction, use ofvasopressors, and use of corticosteroids were left to the treatingphysician's discretion.

2.3. Study outcomes

The primary study outcome was the need for vasopressor supportwithin 72 hours postintubation. Secondary outcome measuresincluded incidence of postintubation hypotension, use of multiplevasopressors within 72 hours postintubation, total duration of

vasopressor use, time to vasopressor initiation, time to vasopressorwithdrawal, duration of mechanical ventilation, ICU and hospital LOS,and all-cause mortality.

2.4. Definitions

The criteria for sepsis and severe sepsis followed those defined bythe American College of Chest Physicians/Society of Critical CareMedicine Consensus Conference [18,19]. Adrenal insufficiency wasdetermined by nonresponse to corticotropin stimulation test asdefined by an increase in serum cortisol level of less than 9 μg/dL ora random cortisol level of less than 15 μg/dL [8,10]. A health care–associated infection was defined by the American Thoracic Society/Infectious Diseases Society of America [20]. Community-acquiredinfection was defined as an infectious process that failed to meet thecriteria for health care–associated infection. Initiation of appropriateempirical antibiotics was defined by the start of antimicrobialregimens for which presumptive or definitive pathogens weresusceptible to in vitro. In the case of culture-negative sepsis or severesepsis, broad spectrum antibiotics were deemed appropriate based onthe presumptive source of infection and in accordance with localpractice guidelines [21]. Indications for intubation included respira-tory compromise, hemodynamic collapse, altered neurologic status(eg, inability to protect the airway), or trauma as documented inelectronic or paper chart intubation notes. Hypotension was definedby a MAP decrease of more than 40% and MAP of less than 70 mm Hg,an MAP of less than 60 mm Hg, or SBP of less than 90 mm Hg lastingmore than 15 minutes despite 500- to 1000-mL intravenous fluidbolus, initiation of vasopressor, or increment of more than 30% of thevasopressor dose [14]. Use of multiple vasopressors was defined bythe initiation of more than 1 vasopressor with at least 1 vasopressorrunning at the maximum infusion rate within 72 hours postintuba-tion. Maximum infusion rates were as follows: epinephrine, 1 μg/kgper minute; norepinephrine, 3.3 μg/kg per minute; phenylephrine,9 μg/kg per minute; dopamine, 20 μg/kg per minute; or vasopressin,0.1 U/min. The time to vasopressor initiation was determined as thetime from intubation to initiation of the first vasopressor, whereas thetime to vasopressor withdrawal was defined as the time fromintubation to discontinuation of all vasopressors for at least 24 hours.

2.5. Statistical analysis

Descriptive statistics were used to report demographic and otherbaseline characteristics. χ2 Test or Fisher exact test was used toanalyze categorical data. Continuous data were analyzed using theStudent t test for parametric data, whereas Wilcoxon rank sum testwas used for nonparametric continuous data. Logistic regression wasused to generate a propensity score (logit) for each subject. Thisscore was matched 1:1 between those receiving and not receivingetomidate using a random seed, nearest neighbor (caliper ≤0.25),without replacement [22]. Results are reported for continuousvariables as the median (interquartile range) for nonparametricdata or mean ± SD for parametric data and for categorical variablesas frequency distributions. A 2-sided P b .05 was consideredstatistically significant.

The study sample size was calculated to achieve at least 80% powerto detect a 25% difference in vasopressor use between the groupsusing an α value of .05. Approximately 300 etomidate and 100nonetomidate patients were needed to match 70 patients betweenthe groups. Patients who died or were discharged within 72 hourswere included in this intention-to-treat analysis.

Propensity score matching was used to adjust for baselinedifferences and minimize bias between patients who receivedetomidate for intubation and patients who did not receive etomidate.Variables included in the propensity score calculation were age, sex,weight, indication for intubation, location of intubation, APACHE II

519N.J. Alday et al. / Journal of Critical Care 29 (2014) 517–522

score, GCS score, SOFA score, baseline MAP, and diagnosis of severesepsis. Logistic regression was used to generate the propensity scorefor each patient, and it was calibrated using the Hosmer-Lemeshowgoodness-of-fit test. Standardized differences were used to assesscovariate imbalances between the unmatched and propensity-matched cohorts. All statistical analyses and propensity scoregeneration and matching were run using Stata (version 11.0; StataCorporation, College Station, TX).

3. Results

3.1. Results of unmatched analysis

Between January 1, 2007, and October 31, 2012, there were 2587patients identified with ICD-9 codes for mechanical ventilation andsepsis. We excluded 2176 patients, leaving a total of 411 patients forevaluation. Etomidate was administered to 303 patients, whereas theremaining 108 patients did not receive etomidate (Fig. 1). The mostcommon reasons for exclusion were intubation before admission,history of chronic immunosuppression, or vasopressor use within 24hours of intubation. Baseline characteristics for the entire cohort aresummarized in Table 1. Overall, 52.1% of patients were male, 63% ofpatients were white, and the mean age was 60 ± 16 years. Mostpatients (93.9%) fulfilled the criteria for the diagnosis of severe sepsis,with health care–associated infection suspected in 74.9% of patients.

The incidence of clinical hypotension was statistically significantbetween the unmatched groups (82.8%; 70.4%; P = .006). Althoughthere were similar proportions of patients in the unmatched cohortswho received vasopressors (59.1%; 51.9%; P = .19; Fig. 2), patientsin the etomidate group required significantly higher doses ofnorepinephrine than those in the nonetomidate group (0.3 vs 0.15μg/kg per minute; P = .02). Those who received etomidate were alsomore likely to receive multiple vasopressors (13.5%; 6.0%; P = .05).Overall, the in-hospital all-cause mortality of the entire cohort was40.6%, the mean ICU LOS was 10.6 ± 9.1 days, the mean hospital LOSwas 16.8 ± 12.7 days, and the median duration of mechanicalventilation was 5.4 (2.5-10) days, with no differences observedbetween the unmatched groups.

Fig. 1. Patients identified, analyzed, exclu

3.2. Results of propensity-matched analysis

Logistic regression was used to generate the propensity score andwas adequately calibrated based on a nonsignificant Hosmer-Lemeshow goodness-of-fit P value of .44. The model had gooddiscrimination based on area under the receiver operator character-istic curve of 0.79. A total of 166 (83 etomidate, 83 nonetomidate)patients were matched based on propensity score. After propensityscore matching, the 2 groups were balanced with regard to patientdemographics and intubation characteristics, including severity ofillness before intubation, baseline MAP, location of intubation, andindication for intubation. Most patients were intubated in the medicalICU, general medical/surgical floor, or emergency department. Themost frequent source of infection was pulmonary, and the mostcommon reason for intubation was respiratory distress. At baseline,the groups were comparable with regards to the absence of significanthypotension, with the median baseline MAP reported as 84 (69-101)mm Hg in the etomidate group and 81 (69-102.6) mm Hg in thenonetomidate group (P = .86). Serum cortisol levels were drawn in21 patients (25.3%) in each group, withmedian baseline cortisol levelssimilar between groups. Corticosteroids were administered to 22(26.5%) etomidate and 16 (19.3%) nonetomidate patients afterintubation, with a median duration of use of 2.6 (1.2-7.5) days and2.2 (0.8-3.2) days, respectively (Table 1).

After propensity score matching, 69 patients (83.1%) in theetomidate group and 63 patients (75.9%) in the nonetomidate groupexperienced significant hypotension (odds ratio [OR], 1.56; 95%confidence interval [CI], 0.73-3.36; P= .25), with the lowest reportedMAP values comparable between groups (Fig. 3). There was nosignificant difference between the 2 matched groups in the primaryoutcome, with 46 (55.4%) etomidate patients and 47 (56.6%)nonetomidate patients requiring vasopressors (OR, 0.95; 95% CI,0.52-1.76; P = .88; Fig. 2). The median duration of vasopressors was1.7 (0.9-3) days in the etomidate group and 1.9 (1-3.5) days in thenonetomidate group. There was also no difference in the maximumnorepinephrine infusion rate or the need for multiple vasopressors(OR, 1.22; 95% CI, 0.36-4.15; P = .75). There were no significantdifferences detected between propensity-matched groups in anysecondary outcomes (Table 2).

ded, and propensity score matched.

Table 1Unmatched and propensity score–matched baseline characteristics

Unmatched cohort Propensity-matched cohort

Etomidate (n = 303) Nonetomidate (n = 108) P STD Etomidate (n = 83) Nonetomidate (n = 83) P STD

Male, n (%) 154 (50.8) 59 (54.6) .49 −0.08

50 (60.2) 45 (54.2) .43 0.12

Age, ya 59.7 ± 14.6 60.8 ± 19.3 .54 −0.07

60.8 ± 14.3 59.6 ± 15.6 .61 0.08

Weight, kgb 80 (64.2-101) 83 (71.9-109.8) .12 −0.19

90.7 ± 26.1 88.2 ± 30.3 .57 0.09

APACHE IIa 21.8 ± 6.6 21.7 ± 6.8 .82 0.03 22 ± 6.3 21.8 ± 6.9 .81 0.04SOFAa 6.3 ± 2.6 6 ± 2.5 .39 0.1 6.7 ± 2.5 6.3 ± 2.4 .29 0.17GCSb 13 (9-15) 13 (9-15) .93 −

0.0313 (9-15) 12 (8-15) .39 0.11

Severe sepsis, n (%) 280 (92.4) 106 (98.1) .032 −0.24

83 (100) 82 (98.8) N .99 0.16

Source of sepsis, n (%)Pulmonary 186 (61.4) 54 (50) .08 40 (48.2) 46 (55.4) .21Gastrointestinal 36 (11.9) 13 (12) 16 (19.3) 10 (12)Renal/GU 34 (11.2) 15 (13.9) 12 (14.5) 8 (9.6)Vascular 16 (5.3) 12 (11.1) 7 (8.4) 11 (13.3)Other 31 (10.2) 14 (13) 8 (9.6) 8 (9.6)

Health care associated, n (%) 221 (72.9) 85 (78.7) .24 62 (74.7) 65 (78.3) .58Appropriate antibiotic regimen, n (%) 296 (97.7) 105 (97.2) .79 81 (97.6) 83 (100) .49Intubation location, n (%)MICU 104 (34.3) 30 (27.8) .07 26 (31.3) 24 (28.9) .74ED 117 (38.6) 38 (35.2) 31 (37.3) 33 (39.8)General ward 44 (14.5) 20 (18.5) 15 (18.1) 13 (15.7)SICU 13 (4.3) 7 (6.5) 4 (4.8) 6 (7.2)Other 25 (8.3) 13 (12) 7 (8.4) 7 (8.4)

Intubation indication, n (%)Respiratory 233 (76.9) 81 (75) .59 63 (75.9) 62 (74.7) .71Neurologic 45 (14.9) 17 (15.7) 12 (14.5) 13 (15.7)Hemodynamic 18 (5.9) 5 (4.6) 7 (8.4) 5 (6)Other 7 (2.3) 5 (4.6) 1 (1.2) 3 (3.6)

Etomidate dose, mg/kga 0.25 ± 0.1 N/A N/A 0.22 ± 0.08 N/A N/AAlternative/adjunctive induction agents, n (%)Midazolam 82 (27.1) 35 (32.4) .29 31 (37.3) 29 (34.9) .59Propofol 8 (2.6) 20 (18.5) .0001 2 (2.4) 15 (18.1) .001Ketamine 0 5 (4.6) 0 5 (6) .059Fentanyl 11 (3.6) 5 (4.6) .001 4 (4.8) 3 (3.6) N .99Neuromuscular blocker 152 (50.2) 36 (33.3) .77

.00343 (51.8) 26 (31.3) .007

Baseline MAP, mm Hgb 80 (69-97) 79.5 (68-120) .99 −0.06

84 (69-101) 81 (69-102) .87 0

STD indicates standardized differences (etomidate − nonetomidate): difference in means or proportions divided by pooled SD; N0.1 chosen as priori to indicate imbalance; GU,genitourinary; MICU, medical intensive care unit; ED, emergency department; SICU, surgical intensive care unit.

a Data are presented as mean ± SD.b Data are presented as median (25%-75% interquartile range).

520 N.J. Alday et al. / Journal of Critical Care 29 (2014) 517–522

4. Discussion

In this study, we evaluated the effect of etomidate use in patientswith sepsis or severe sepsis and found no differences in vasopressoruse within the 72 hours after administration for intubation.Furthermore, we found no statistical differences in any secondary

Fig. 2. Vasopressor requirements within 72 hours postinduction.

outcomes, including incidence of significant hypotension, ventilatordays, ICU and hospital LOS, and mortality.

The safety of single-bolus etomidate for intubation in septicpatients continues to be debated. Results reported from earlier studiesraised concerns about the potential clinical consequences of the

Fig. 3. Mean arterial pressure before induction agent and up to 72 hours postinductionagent administration.

Table 2Unmatched and propensity-matched outcomes

Unmatched cohort Propensity-matched cohort

Etomidate (n = 303) No etomidate (n = 108) P STD Etomidate (n = 83) No etomidate (n = 83) P STD

Hypotension, n (%) 251 (82.8) 76 (70.4) .006 0.31 63 (75.9) 69 (83.1) .25 0.18Norepinephrine max dose, μg/kg/mina 0.3 (0.12-0.8) 0.15 (0.09-0.56) .02 0.23 0.42 (0.1-1) 0.16 (0.1-0.6) .09 0.27Multiple vasopressors, n (%) 41 (13.5) 7 (6.5) .05 0.22 6 (7.2) 5 (6) .76 0.05Duration of vasopressor, da 2 (0.9-3.3) 1.8 (1-3) .81 −0.37 1.7 (0.9-3) 1.9 (1-3.5) .69 −1.18Cortisol level measured, n (%) 76 (25.1) 25 (23.1) .69 0.05 21 (25.3) 21 (25.3) N .99 0Baseline serum cortisol, μg/dLa 22.9 (15.6-36.5) 23.4 (18.4-35.9) .71 0.01 21.6 (16-36.1) 23.2 (18.4-35.9) .65 −0.02Postintubation corticosteroid use, n (%) 88 (29) 19 (17.6) .02 0.26 22 (26.5) 16 (19.3) .27 0.17Duration of corticosteroid use, da 4 (1.8-8) 2.2 (1-5) .09 0.15 2.6 (1.2-7.5) 2.2 (0.8-3.2) .49 −0.14Duration of mechanical ventilation, da 5 (2.2-10) 6 (3-10.8) .27 −0.24 5.4 (2-10.4) 6 (3-10) .38 −0.17ICU LOS, da 8 (4-13) 9 (5-14.8) .22 −0.16 7 (3.9-13) 9 (5-15) .19 −0.17Hospital LOS, da 13 (8-21) 13.6 (9-23) .19 −0.16 14 (7-23) 13.5 (9-23) .76 −0.05All-cause mortality, n (%) 125 (41.3) 42 (38.9) .67 0.05 38 (45.8) 36 (43.4) .76 0.05

Data are presented as mean ± SD.a Data are presented as median (25%-75% interquartile range).

521N.J. Alday et al. / Journal of Critical Care 29 (2014) 517–522

transient period of etomidate-associated adrenal insufficiency [8-11].A retrospective cohort study including patients with severe sepsis andseptic shock investigated the association between adrenal dysfunc-tion and mortality and found that etomidate was associated with anincreasedmortality risk (OR, 1.53; 95% CI, 1.1-2.3), although statisticalsignificance was lost after adjustment for severity of illness in themultivariate analysis. In both univariate and multivariate analyses,vasopressor requirements were also significantly associated with anincreased risk of mortality (OR, 38.5 [95% CI, 20.7-71.7]; OR, 105.8[95% CI, 25.1-446.9]) [10]. More recently, a meta-analysis of 5randomized controlled trials evaluated the risk of mortality associatedwith etomidate administration in septic patients. In the 865 patientsincluded, those who received etomidate had a higher rate of mortality(relative risk, 1.2; 95% CI, 1.02-1.42) [12]. However, in anotherretrospective study, the difference in hospital mortality betweensevere sepsis and septic shock patients who received etomidate(43.9%) and those who did not receive etomidate (45.6%, P = .87)failed to attain significance [13].

It remains unclearwhether significant hemodynamic instability mayresult frometomidate-associated adrenal insufficiency and contribute tothe development of adverse clinical outcomes in septic patients. In asmall retrospective cohort study comparing etomidate against metho-hexital, 61% patients in the methohexital group and 74% patients in theetomidate group experienced clinical hypotension (P = .53), with 10patients (43%) in the methohexital group compared with 12 patients(52%) in the etomidate group requiring a more than 30% increase invasopressor dose after intubation [14]. In our study, patients whoreceived etomidate required median norepinephrine doses 2.6 timeshigher (P = .09) than patients who did not receive etomidate, despitesimilar rates of clinical hypotension between groups. However, theclinical significance of this effect was not reflected in a difference inmortality rates between groups. Our study included patients that werehemodynamically stable at baseline with median MAPs of 80 mmHg orhigher. Furthermore, of the 25%of patientswithmeasured cortisol levels,there were no patients who met our definition of adrenal insufficiency.Patients included in this study may have been less susceptible to thedevelopment of clinically significant adrenal insufficiency and thesubsequent potential for adverse cardiovascular outcomes.

Although several studies have demonstrated adverse outcomesassociated with etomidate use, our study showed no statisticaldifferences in primary or secondary outcomes. Our results are similarto previously published studies that showed no harm associated withetomidate when used for intubation in septic patients. A retrospectivecohort study reported no significant differences in mortality, ICU LOS,or vasopressor use within 72 hours of intubation in patients withsevere sepsis or septic shock who received etomidate and patientswho did not receive etomidate [7]. A randomized controlled trial by

Jabre et al [15] comparing etomidate against ketamine for induction incritically ill patients reported a greater proportion of patients withadrenal insufficiency who received etomidate compared with keta-mine (OR, 6.7; 95% CI, 3.5-12.7), althoughmortality rates did not differsignificantly (31% [95% CI, 23-39] vs 21% [95% CI, 13-29], respectively;P = .11). In a subgroup analysis including only septic patients (n =76), no significant differences in mortality between the groups werefound. Another retrospective study in septic shock patients failed toshow any statistical difference between induction agent (etomidate,propofol, thiopental, other agent, no agent), vasopressor, inotrope,steroid use, or mortality [16].

There are inherent limitations associated with a retrospective,observational study design. We relied on ICD-9 diagnosis codes toselect patients who required mechanical ventilation and who werediagnosed with sepsis or severe sepsis, which may not have capturedall potentially eligible subjects. The selection of induction agents andthe clinical decision making regarding intubation, extubation, admin-istration and dose titration of vasopressors, and monitoring were atthe discretion of the physician. Patients who received corticosteroidspostinductionwere not excluded. However, a recent study by Payen etal [23] analyzed the effects of moderate-dose hydrocortisone onvasopressor requirements 48 hours postetomidate administrationand found no statistical difference when compared with saline (0.9%NaCl). Hydrocortisone had no impact on the need for vasopressors butdid decrease the vasopressor dose at a significantly higher rate thanplacebo. In our study, 26.5% of patients in the etomidate groupreceived postintubation corticosteroids compared with 19.3% ofpatients who did not receive etomidate.

Our study has several strengths and differs from previous studiesbecause itwas amulticenter trial and used propensity scorematching toaccount for potential confounders, such as severity of illness andbaseline MAP, and minimize bias between the groups. We used ahemodynamic variable as our primary outcome and only includedpatientswith sepsis or severe sepsis. Previous studieswere not designedto evaluate hemodynamic outcomes, whichwere reported as secondaryor post hoc analyses and included patients who developed septic shockbefore receiving etomidate or another induction agent for intubation.

5. Conclusion

In this multicenter, retrospective propensity-matched study, the useof etomidate for intubation in patients with sepsis or severe sepsis didnot increase vasopressor requirementswithin 72hours after intubation.Etomidate was not associated with significant differences in duration ofmechanical ventilation, ICU or hospital LOS, or all-cause mortality.Single-dose etomidate for intubation in patients with sepsis or severesepsis should be used judiciously in consideration for the potential

522 N.J. Alday et al. / Journal of Critical Care 29 (2014) 517–522

adverse effects associated with adrenal insufficiency. However, furtherprospective studies should be performed to evaluate the hemodynamicconsequences associated with etomidate in septic patients.

References

[1] Amidate (etomidate) injection [package insert]. Lake Forest, IL: Hospira Inc.;November 2009.

[2] JacksonWL. Should we use etomidate as an induction agent in patients with septicshock. Chest 2005;127:1031–8.

[3] Forman SA. Clinical and molecular pharmacology of etomidate. Anesthesiology2011;114:695–707.

[4] Absalom A, Pledger D, Kong A. Adrenocortical function in critically ill patients 24 hafter a single dose of etomidate. Anaesthesia 1999;54:861–7.

[5] Bloomfield R, Noble DW. Etomidate and fatal outcome—even a single bolus dosemay be detrimental for some patients. Br J Anaesth 2006;97:116–7.

[6] Vinclair M, Broux C, Faure P, Brun J, Genty C, Jacquot C, et al. Duration of adrenalinhibition following a single dose of etomidate in critically ill patients. IntensiveCare Med 2008;34:714–9.

[7] Dmello D, Taylor S, O’Brien J, Matuschak GM. Outcomes of etomidate in severesepsis and septic shock. Chest 2010;138:1327–32.

[8] Annane D, Sebille V, Charpentier C, Bollaert P, Francois B, Korach J, et al. Effect oftreatment with low doses of hydrocortisone and fludrocortisones on mortality inpatients with septic shock. JAMA 2002;288(7):862–71.

[9] Sprung CL, Annane D, Keh D,Moreno R, SingerM, Freivogel K, et al. Hydrocortisonetherapy for patients with septic shock. N Engl J Med 2008;358:111–24.

[10] Lipiner-Friedman D, Sprung CL, Laterre PF, Weiss Y, Goodman S, Vogeser M, et al.Adrenal function in sepsis: the retrospective CORTICUS cohort study. Crit CareMed 2007;35:1012–8.

[11] Cuthbertson BH, Sprung CL, Annane D, Chevret S, Garfield M, Goodman S, et al. Theeffects of etomidate on adrenal responsiveness and mortality in patients withseptic shock. Intensive Care Med 2009;35:1868–76.

[12] Chan CM, Mitchell AL, Shorr AF. Etomidate is associated with mortality andadrenal insufficiency in sepsis: a meta-analysis. Crit Care Med 2012;40:2945–53.

[13] Ehrman R, Wira C, Lomax A, Hayward A, Marcelin J, Ellis T, et al. Etomidate use insevere sepsis and septic shock patients does not contribute to mortality. InternEmerg Med 2011;6:253–7.

[14] Diaz-Guzman E, Mireles-Cabodevila E, Heresi GA, Bauer S, Arroliga A. Acomparison of methohexital versus etomidate for endotracheal intubation ofcritically ill patients. Am J Crit Care 2010;19:48–54.

[15] Jabre P, Combes X, Lapostolle F, Dhaouadi M, Ricard-Hibon A, Vivien B, et al.Etomidate versus ketamine for rapid sequence intubation in acutely ill patients: amulticentre randomised controlled trial. Lancet 2009;374:293–300.

[16] Ray DC, McKeown DW. Effect of induction agent on vasopressor and steroid useand outcome in patients with septic shock. Crit Care 2007;11:R56.

[17] Annane D. Etomidate and intensive care physicians. Intensive Care Med 2005;31(10):1454.

[18] Bone RC, Balk RA, Cerra FB, Dellinger R, Fein A, KnausW, et al. Definitions for sepsisand organ failure and guidelines for the use of innovative therapies in sepsis. TheACCP/SCCM Consensus Conference Committee. American College of ChestPhysicians/Society of Critical Care Medicine. Chest 1992;101(6):1644–55.

[19] Dellinger R, Levy M, Carlet J, Bion J, Parker M, Jaeschke R, et al. Surviving SepsisCampaign: international guidelines for management of severe sepsis and septicshock: 2008. Crit Care Med 2008;36(1):296–327.

[20] American Thoracic Society, Infectious Diseases Society of America. Guidelinesfor the management of adults with hospital-acquired, ventilator-associated, andhealthcare-associated pneumonia. Am J Respir Crit Care Med 2005;171:388–416.

[21] Gaieski DF, Mikkelsen ME, Band RA, Pines J, Massone R, Furia F, et al. Impact oftime to antibiotics on survival in patients with severe sepsis or septic shock inwhom early goal-directed therapy was initiated in the emergency department.Crit Care Med 2010;38(4):1045–53.

[22] Leuven E, Sianesi B. PSMATCH2: Stata module to perform full Mahalanobis andpropensity score matching, common support graphing, and covariate imbalancetesting. Statistical Software Components S432001, Boston College Department ofEconomics; 2003 [revised 19 Apr 2012. Available at: http://ideas.repec.org/c/boc/bocode/s432001.html. Accessibility verified June 12, 2012].

[23] Payen J,DupuisC, Trouve-BuissonT, VinclairM,BouzatP, GentyC, et al. Corticosteroidafter etomidate in critically ill patients: a randomized controlled trial. Crit Care Med2012;40(1):29–35.