Sepsis is a life-threatening syndrome caused by a dysregulated host response to infection that may progress to fatal shock and results in greater than 5 million annual deaths worldwide (1,2). Prioritizing timely interventions to restore adequate perfusion have been demonstrated to improve morbidity and mortality (3–6). Guideline recommendations for the early treatment of septic shock recommend 30 mL/kg of fluid resuscitation followed by additional fluid or vasopressor therapy if initial fluid resuscitation is inadequate to restore and maintain cellular perfusion (7). Although vasopressor therapy is commonly used in the treatment of septic shock, there is wide variation in the selection, sequence of initiation, and titration of vasoactive agents (4,8–13). Additionally, there is increasing discussion that individual patients may benefit from a tailored vasopressor treatment strategy (14).
Although fluids and vasopressors are recognized as essential components of resuscitation (7), the optimal timing and dosing of each treatment for patients with septic shock remain unknown. Substantial data have linked excessive fluid administration to worse outcomes (e.g., acute kidney injury, days in organ failure, mortality) (15–22), leading some investigators to explore early fluid minimization strategies, in particular the early, and possibly preferential use of vasopressors (13,23). However, excessive vasopressor exposure has also been linked to worse outcomes, and there are several potential adverse effects that may be vasopressor dosage related (e.g., arrhythmia, splanchnic perfusion deficits) (24–27). Furthermore, it is plausible that the effects of the two resuscitation treatments interact. That is, the effect of vasopressors on hemodynamics and outcome depends on the fluid status at the time of vasopressor administration. Few studies have examined this important question.
The goals of this study were to characterize vasopressor therapy in the early phases of septic shock resuscitation and to determine its association with mortality. Given the critical time-dependence of sepsis therapies, we focused analysis on the first 6 and 24 hours of sepsis resuscitation. We quantified vasopressor dosing intensity (VDI) as the total vasopressor dose infused across all vasopressors in norepinephrine equivalents (NEE). Using this definition, we aimed to 1) determine the association between VDI during the first 6 hours and first 24 hours after the onset of septic shock and 30-day in-hospital mortality; 2) determine whether the effect of VDI varies depending on fluid resuscitation volume; and 3) determine whether the effect of VDI varies depending on the pattern of dosing titration.
MATERIALS AND METHODS
CHaracterizAtion of vaSoprEssor Requirements in Shock (CHASERS) study is a prospectively defined substudy of Observation of VariatiOn in fLUids adMinistEred in shock-CHaracterizAtion of vaSoprEssor Requirements in Shock (VOLUME-CHASERS) a multicenter, prospective, observational study, conducted through the Discovery Network, the Society of Critical Care Medicine’s research network (ClinicalTrials.gov: NCT03190408) (Protocol S1, Supplemental Digital Content 1, http://links.lww.com/CCM/F610). From September 1, 2017, to February 1, 2018, consecutive adults (≥ 18 yr) admitted to a United States (n = 32) or Jordanian (n = 1) study hospital with shock were enrolled over a 2–4-week period. Patients could be enrolled from any floor within the institution if there were plans to transfer to an ICU. Shock was defined as meeting one of the following criteria: 1) need for vasopressor therapy (at any infusion dose) to maintain a mean arterial pressure (MAP) greater than 65 mm Hg or 2) systolic blood pressure less than 90 mm Hg, with shock onset defined as the date and time of when the first criterion was met. Patients were excluded if they: 1) were in the operating room at the time of shock onset; 2) were admitted to an ICU after cardiac surgery with cardiogenic shock as the primary etiology of shock; 3) were transferred from another hospital or emergency department (ED) to a study hospital; or 4) were previously enrolled in the study. Patients in the CHASERS substudy were restricted to those with septic shock that were treated with appropriate initial antibiotics and greater than or equal to 1 vasopressor within 24 hours after shock onset. The diagnosis of septic shock was made at each study site by treating clinicians in accordance with the Surviving Sepsis Campaign international guideline definitions (7). CHASERS’ patients were additionally excluded if their resuscitation involved mechanical support (e.g., extracorporeal life support) and if there was missing vasopressor dose or covariate data. This study was approved by the Institutional Review Board at all 33 participating hospitals. The requirement for informed consent was waived by all centers. No external funding was provided for this study.
We collected baseline demographics, shock etiology, medication history, comorbidities, location of shock onset, and ICU type. Variables needed to calculate the Acute Physiology and Chronic Health Evaluation (APACHE) III and Sequential Organ Failure Assessment (SOFA) scores were collected during the time period from 12 hours before to 12 hours after shock onset. During the 24-hour period following shock onset, summary measures of resuscitation treatments were collected for the following time periods: hour 0–3, 3–6, 6–12, and 12–24. Fluid administration was collected as the total volume administered during each time period (crystalloid, colloid, packed RBCs, fresh frozen plasma, and platelets). Vasopressor administration was collected as the maximum and minimum infusion dose administered during each time period (dopamine, epinephrine, norepinephrine, phenylephrine, and vasopressin). Mechanical ventilation was collected as being ventilated versus not ventilated, for any length of time, during each time period. The use of renal replacement therapy during the first 24 hours was collected. Patients were followed until hospital discharge or death, whichever came first. In-hospital mortality was determined through medical record review. Training on data collection procedures was provided by the study’s executive committee to each site. De-identified site data were uploaded into a secure online form using Research Electronic Data Capture (REDCap) (https://project-redcap.org) electronic data capture tools (28).
The primary exposure was VDI, quantified as the total vasopressor dose infused across all vasopressors in NEE and expressed in μg/min (8,29). We chose μg/min (as opposed to μg/kg/min) because the majority of study centers (21/33, 64%) used μg/min for norepinephrine dosing. VDI was determined by first taking the average of the lowest and highest infusion dose documented for each vasopressor at each time period, and then calculating the corresponding NEE for that time period for each vasopressor administered. This calculation assumes that the patient was maintained at a constant infusion dose midway between the lowest and highest infusion doses over each time period. We assumed a flat dose of 0.04 U/min for intervals where vasopressin was administered. Once VDI was obtained for each time period (hour 0–3, 3–6, 6–12, and 12–24), we calculated the cumulative average VDI at 6 and 24 hours by taking a time-weighted average of VDI from 0 to 6 hours and 0–24 hours. Similarly, the cumulative volume of fluid administration was tabulated for the time periods from 0 to 6 hours and 0–24 hours. In a secondary analysis that aimed to examine the timing of vasopressor titration, we categorized VDI for the time periods of 0–6 and 0–24 hours as low-dose (VDI < 15 μg/min NEE) or high-dose (VDI ≥ 15 μg/min NEE) (8). We then classified the VDI across the two time periods into four categories: 1) “never high” (VDI was < 15 μg/min NEE for both time periods); 2) “early high and late low“ (VDI was ≥ 15 μg/min NEE during the first 6 hr but titrated < 15 μg/min NEE by the end of the 24 hr period); 3) “early low and late high” (VDI was < 15 μg/min NEE for the first 6 hr but titrated ≥ 15 μg/min NEE by the end of the 24 hr period); and 4) “always high” (VDI was ≥ 15 μg/min NEE during the first 6 hr and remained elevated for the entire 24 hr period).
The association between VDI and in-hospital mortality at 30 days was modeled using multivariable logistic regression, with separate models constructed for VDI at 0–6 hours and VDI at 0–24 hours. Each model examined potential interaction between VDI and fluid administration by testing interaction terms using the likelihood ratio test. Significant interactions were visualized with plots of the population average marginal mortality estimates for various combinations of fluid volume and VDI (30). In a secondary analysis, an additional model was constructed to examine the association between VDI titration and in-hospital mortality at 30 days. All models were adjusted for potential confounding variables that were selected a priori based on clinical knowledge and prior literature (31): age, sex, APACHE III score, need for mechanical ventilation, corticosteroid administration, location of shock onset, hospital length of stay prior to shock onset, treating ICU type, and past medical history (heart failure, diabetes mellitus, renal disease [with and without dialysis], liver disease, metastatic cancer, and hematologic malignancy). To account for the potential effects of body size, all models adjusted for height and weight.
We adjusted for potential center effects by repeating all models using random-effects logistic regression (32,33), with center specified as a random intercept. To determine whether results were sensitive to our choice of vasopressor dosing unit, we repeated analyses using weight-based VDI (μg/kg/min) (34). Because data on baseline MAP and fluid administration was not available for all subjects, these variables were not included in the primary analysis models. However, we repeated all analyses in the subset of patients with baseline fluid administration data available (n = 386), with baseline MAP (n = 476), and with both variables (n = 385). Last, we repeated the analysis after restricting patients surviving the initial 24-hour period (n = 607) to evaluate potential survival bias. All statistical analyses were performed using STATA release 14.2 (StataCorp LLC, College Station, TX).
The VOLUME-CHASERS study enrolled 1,639 patients, and after excluding patients without septic shock or vasopressor therapy, the CHASERS study included 616 patients (Fig. S1, Supplemental Digital Content 1, http://links.lww.com/CCM/F610). Baseline demographics are reported in Table 1 and Table S1 (Supplemental Digital Content 1, http://links.lww.com/CCM/F610). Patients were most often diagnosed with septic shock in the ED, had a high severity of illness (APACHE III score 97 [± sd 28]), and were most often admitted to a medical or mixed medical/surgical ICU. In-hospital mortality at 30 days was 31%.
TABLE 1. -
Baseline Demographics and Outcomes
|Age, yr, mean (sd)
|Weight, kg, mean (sd)a
|Male, n (%)
|Race, n (%)
| African American
| Not reported
|Acute Physiology and Chronic Health Evaluation III, mean (sd)
|Sequential Organ Failure Assessment, median (IQR)
|Lactate, mg/dL, median (IQR)
| Highest in the perishock periodb,c
| Highest in the 0–24 hr after shock onsetd
|Location at start of shock, n (%)
| Emergency department
| Postanesthesia care unit
|ICU type, n (%)
|LOS prior to shock onset, d, n (%)
| > 14
|Mechanical ventilation within first 24 hr, n (%)
|ICU LOS, d, median (IQR)a
|Hospital LOS, d, median (IQR)a
|In-hospital mortality at 30 d, n (%)e
IQR = interquartile range, LOS = length of stay.
an = 3 missing values.
bn = 50 missing values.
cHighest lactate in the perishock period is defined as the highest lactate in the 12 hr before and 12 hr following shock onset.
dn = 52 missing values.
eThe number of patients who died within 24 hr was nine of 616 (1.5%). The number of patients who died within 48 hr was 30 of 616 (4.9%).
Vasopressor and Fluid Administration Data
Table 2 reports vasopressor and fluid data. Patients received a median of 1,160 mL (interquartile range [IQR], 548–2,226 mL) of IV fluid during the 12 hours before shock onset and a median of 3,400 mL (IQR, 1,851–5,338 mL) during the 24 hours after shock diagnosis. Individual fluid amounts are reported in Table S2 (Supplemental Digital Content 1, http://links.lww.com/CCM/F610). Crystalloids were the predominant fluid administered at all time periods. The most common vasopressor was norepinephrine (93%), followed by vasopressin (39%), phenylephrine (18.5%), epinephrine (10%), and dopamine (6%). The number of vasopressors administered increased over time within the first 24 hours after shock onset (Fig. 1). The median vasopressor infusion dose administered during the first 24 hours was 8.5 μg/min NEE (IQR, 3.4–18.1 μg/min NEE). Individual vasopressor infusion doses for each period are reported in Table S2 (Supplemental Digital Content 1, http://links.lww.com/CCM/F610). High-dose vasopressor support was required in 31% of patients. Compared to low-dose vasopressor support, the high-dose group had a higher APACHE III score, SOFA score, and mortality (Table S3, Supplemental Digital Content 1, http://links.lww.com/CCM/F610). MAP was consistently above 65 mm Hg during the first 24 hours after shock in both the high-dose and low-dose groups (Fig. S2, Supplemental Digital Content 1, http://links.lww.com/CCM/F610).
TABLE 2. -
, and Ancillary Medication Administration Data
||Result (n = 616)
| Vasopressor use, n (%)
| Total vasopressor amount (mg NEE)
| 0–24 hr
| Vasopressor infusion dose (μg/min NEE)
| 0–3 hr, n = 402
| 3–6 hr, n = 454
| 6–12 hr, n = 509
| 12–24 hr, n = 517
| 0–24 hr, n = 616
| High vasopressor infusion dose use (rate ≥ 15 μg/min NEE)
| Preshock fluid (12 hr prior to shock onset), mL, n = 373
| Total fluidc,d, mL
| 0–3 hr, n = 479
| 3–6 hr, n = 427
| 6–12 hr, n = 488
| 12–24 hr, n = 495
| Total over 24 hr, n = 582
| Hydrocortisone use (≥ 200 mg/d), n (%)
| Propofol or dexmedetomidine use, n (%)
| 0–3 hr
| 3–6 hr
| 6–12 hr
| 12–24 hr
| Ever received within first 24 hr
NEE = norepinephrine equivalents.
aTimeframe is within the first 24 hr of shock onset unless otherwise stated.
bData presented as median (interquartile range) unless otherwise stated.
cCrystalloid plus colloid fluid plus blood product fluid volume.
dBlood product fluid volume consists of packed RBCs, platelets, and fresh frozen plasma.
In the analysis of the 0–6-hour time period, increasing VDI was associated with increased odds of 30-day inpatient mortality (Fig. 2). However, the strength of this association varied depending on the amount of concomitant fluid administration during the 0–6-hour period. As fluid volume increased, the association between VDI and increased mortality was attenuated, such that no significant association between VDI and mortality was observed when the 0–6-hour fluid volume was at least 2,000 mL (Fig. 2; and Tables S4 and S5, Supplemental Digital Content 1, http://links.lww.com/CCM/F610).
In the analysis of the 0–24-hour time period, increasing VDI was associated with increased odds of 30-day inpatient mortality, but this association did not vary depending on fluid resuscitation volume over the 0–24-hour period: every 10 μg/min increase in average VDI over the 24-hour period was associated with 33% increased odds of 30-day mortality (adjusted odds ratio, 1.33; 95% CI, 1.16–1.53) (Table S4 and Fig. S3, Supplemental Digital Content 1, http://links.lww.com/CCM/F610). In the analysis of VDI titration, we observed an association between high-dose vasopressor exposure that varied depending on the titration pattern. In the unadjusted analysis (Fig. 3), patients who exhibited a VDI pattern of early high/late low had a lower associated mortality rate compared with those patients with VDI patterns of early low/late high and sustained high dose. This pattern of results remained after adjustment for potential confounding variables (Table S3, Supplemental Digital Content 1, http://links.lww.com/CCM/F610). Primary and secondary analysis results were similar in random-effects analyses, in those adjusted for baseline MAP and fluid administration, and in those that were restricted to the population surviving the first 24 hours after shock onset (Table S6, Supplemental Digital Content 1, http://links.lww.com/CCM/F610). Last, our results were similar in an analysis based on μg/kg/min rather than μg/min VDI (Table S7, Supplemental Digital Content 1, http://links.lww.com/CCM/F610).
In this multicenter, observational study of septic shock care, we found that increasing VDI during the first 24 hours after the shock was associated with increased mortality risk. Notably, we observed an interaction between VDI and fluid administration during the first 6 hours of shock resuscitation, with the association between VDI and mortality attenuated by increasing volumes of fluid administration. This interaction was no longer apparent when the analysis focused on the entire 24-hour period. We also found that brief, early high VDI exposure was associated with lower mortality compared to later and sustained VDI exposure. Taken together, we hypothesize that early aggressive vasopressor titration may be preferred compared to slower titration to high doses, but that this strategy must be accompanied by adequate fluid resuscitation in the early hours after shock onset to avoid potential deleterious effects of high-dose vasopressors. If true, this hypothesis has important implications, given the recent interest in early vasopressors to prevent excessive fluid resuscitation (9,13,23,35). Although our results suggest that the effects of vasopressors may depend on volume status, the observational nature of our study precludes examination of potential interaction mechanisms, and we consider our results hypothesis-generating only.
Common strategies for vasopressor initiation and dosing include starting norepinephrine at a moderate dose and increasing the dose gradually to achieve a goal MAP. Another common strategy is to administer incremental boluses of fluids while gradually increasing the vasopressor dose to achieve a goal MAP. However, both of these methods may postpone the recognition of “refractory shock” and delay optimization of tissue perfusion. An alternative approach may be to start norepinephrine at a higher dose and reduce the dose after achieving goal MAP, thereby decreasing the time to goal MAP. Further study is needed comparing vasopressor dosing and titration to expand upon our observations.
The harms of excess fluid administration have been extensively described (15–22). As a result, there has been increasing interest in the use of fluid restrictive strategies during the early phases of septic shock, combined with earlier and more aggressive vasopressor support (9,13,23,35). Although such approaches have a physiologic rationale (35), our data suggest that there is some minimum threshold of fluid administration that is required in patients treated with vasopressors during the first hours after shock onset. Our findings are in agreement with previous data that has linked the early administration of vasopressors in lieu of fluid therapy with increased mortality (10). Although our data does not identify a threshold for “adequate fluid resuscitation,” increasing VDI during the first 6 hours was no longer associated with mortality when fluid resuscitation was at least 2 L (~25 mL/kg). Notably, this amount is substantially less than the median fluid volumes in several studies of early goal-directed therapy (5–6 L over 6 hr) (35). Ultimately, achieving an optimal balance between fluids and vasopressors will need to consider both the timing of these interventions and the need to provide individualized therapy.
Our study has limitations. Observational studies are susceptible to confounding, such that patients who required higher VDI exposures had a higher baseline risk for mortality. We addressed this concern by adjusting for an extensive set of potential confounding variables, including age, APACHE III score, need for mechanical ventilation, and comorbid chronic disease. In addition, we conducted multiple sensitivity analyses, with each showing similar results. Given the nature of the study design, it is possible that patients were identified after the onset of septic shock, although reflective of real-world practice. Additionally, patients were followed until hospital discharge, potentially underestimating the effects on longer-term mortality. Our method for estimating vasopressor dosing relied on summary measures for specific time periods, requiring the assumption that vasopressor dose remained constant for up to 12 hours. This creates the potential for measurement error bias. However, VDI was calculated similarly for all patients, and there is no obvious suggestion that any measurement error would be related to baseline mortality risk. Thus, it is plausible that any measurement error would be nondifferential, which would be expected to bias our point estimates toward the null (36). Additional studies that estimate VDI with hour-to-hour data will be required to validate our findings. Of particular interest would be studies that define additional clinically relevant vasopressor dosing patterns. We did not collect positive end-expiratory pressure values or amount of sedation administered which may have affected MAP and subsequent vasopressor dosing. We did not express fluid volumes or vasopressor doses in weight-based units, as we judged that this better reflects clinical practice in most institutions. However, our regression models include height and weight as covariates, thus controlling for body size in all outcome analyses. Finally, NEE may not represent a precise pharmacological equivalency in individual patients. Nevertheless, this approach facilitated a method to describe the intensity of vasopressor therapy and used a dose conversion that has been described in prior investigations (29).
Increasing VDI during the first 24 hours after septic shock was associated with increased mortality. This association varied with the amount of early fluid administration and the timing of vasopressor titration.
The Observation of VariatiOn in fLUids adMinistEred in shock-CHaracterizAtion of vaSoprEssor Requirements in Shock (VOLUME-CHASERS) Study Group acknowledges the following sites and individuals for their efforts on the project. Sites are organized alphabetically.
Bridgeport Hospital: Tina McCurry, BSN; Brigham and Women’s Hospital: Jeremy R. DeGrado, PharmD, Kevin M. Dube, PharmD, Kenneth E. Lupi, PharmD; Cleveland Clinic Foundation: Andrei Hastings, MD, Omar Mehkri, MD; Duke University School of Medicine: Raquel R. Bartz, MD, MMCi, Angela L. Pollak, MD, Sarah Kendall Smith, MD, PhD; Emory University Hospital/Grady Hospital: Marguerite Stewart, Leona Wells; Geisinger Wyoming Valley Medical Center: Jamie Kerestes, PharmD, Kayla Kotch, Sarah Miller; Intermountain Healthcare: Brent Armbruster, Valerie Aston, MBA, Katie Brown, Mardee Merrill; King Hussein Cancer Center: Nadeen Anabtawi, PharmD; Lahey Hospital & Medical Center: Katie Nault, PharmD; Lake Region Medical Center: Kerri L. Federico, PharmD, Peter-John Trapp, PharmD; Mayo Clinic: Joseph C. Farmer, MD, Pablo Moreno Franco, MD, Shurong Gong, Rahul Kashyap, MBBS, Sidhant Singh; University of Texas MD Anderson Cancer Center: Reagan D. Collins, PharmD; Montefiore Medical Center, Albert Einstein College of Medicine: Jorge Ataucuri-Vargas, MD, Vladyslav Dieiev, MD, Ashley Kang, MD, Ann Wang, MD; Mt Sinai Health System: Neha N. Goel, MD, MSCR; New York University: Oscar Mitchell, MD; Ohio Health/Riverside Methodist Hospital: Jordan DeWitt, PharmD, Alex Heine, PharmD, Abby Tyson, PharmD; Oregon Health and Science University: Dubier Matos, Ebaad Haq; Rush University Medical Center: Katie Dalton, PharmD; St Agnes Hospital: Valentina Amaral, MD, Jasmine Aulakh, MD, Nauman Farooq, MD; Truman Medical Center: Kerra Cissne, PharmD; University of Arizona: Jose Camarena, Alexia Demitsas, Kristen Deupree, Karen Lutrick, PhD; University of Cincinnati: Nora Elson, MS, Dina Gomaa, John Shinn III, Anthony Spuzzillo, Devin Wakefield; University of Maryland: Mehrnaz Pajoumand, PharmD, Sharon Wilson, PharmD, Siu Yan, Amy Yeung, PharmD; University of Michigan: Tina Chen, Sinan Hanna; University of Oklahoma Health Sciences Center: Lauren Sinko, Kassidy Malone, Deamber Piel; University of Utah: Chloe Skidmore; University California Los Angeles: Matt Flynn, MD, Ji Yeon Seo, MD; University of Rochester Medical Center: Nicole M. Acquisto, PharmD, Kathryn Connor, PharmD, Samantha Delibert, PharmD, Christine Groth, PharmD, Jeff Huntress, PharmD, Gregory Kelly, PharmD, Therese Makhoul, PharmD, Hannah Mierzwa, Stephen Rappaport, PharmD; University of Southern California: Daisy Rios; Vidant Medical Center: Bethany Crouse, PharmD; Wake Forest Baptist Health: Michael Kenes, PharmD; and Yale-New Haven Health: Shamsuddin Akhtar, MD, Abdalla A. Ammar, PharmD.
1. Fleischmann C, Scherag A, Adhikari NK, et al.; International Forum of Acute Care Trialists: Assessment of global incidence and mortality
of hospital-treated sepsis
. Current estimates and limitations. Am J Respir Crit Care Med 2016; 193:259–272
2. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis
and septic shock
-3). JAMA 2016; 315:801–810
3. Kumar A, Roberts D, Wood KE, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock
. Crit Care Med 2006; 34:1589–1596
4. Beck V, Chateau D, Bryson GL, et al.; Cooperative Antimicrobial Therapy of Septic Shock
(CATSS) Database Research Group: Timing of vasopressor
initiation and mortality
in septic shock
: A cohort study. Crit Care 2014; 18:R97
5. Seymour CW, Gesten F, Prescott HC, et al. Time to treatment and mortality
during mandated emergency care for sepsis
. N Engl J Med 2017; 376:2235–2244
6. Venkatesh B, Finfer S, Cohen J, et al.; ADRENAL Trial Investigators and the Australian–New Zealand Intensive Care Society Clinical Trials Group: Adjunctive glucocorticoid therapy in patients with septic shock
. N Engl J Med 2018; 378:797–808
7. Rhodes A, Evans LE, Alhazzani W, et al. Surviving Sepsis
Campaign: International guidelines for management of sepsis
and septic shock
: 2016. Crit Care Med 2017; 45:486–552
8. Russell JA, Walley KR, Singer J, et al. Vasopressin versus norepinephrine infusion in patients with septic shock
. N Engl J Med 2008; 358:877–887
9. Bai X, Yu W, Ji W, et al. Early versus delayed administration of norepinephrine in patients with septic shock
. Crit Care 2014; 18:532
10. Waechter J, Kumar A, Lapinsky SE, et al.; Cooperative Antimicrobial Therapy of Septic Shock
Database Research Group: Interaction between fluids and vasoactive agents on mortality
in septic shock
: A multicenter, observational study. Crit Care Med 2014; 42:2158–2168
11. Gordon AC, Mason AJ, Thirunavukkarasu N, et al. Effect of early vasopressin vs norepinephrine on kidney failure in patients with septic shock
: The VANISH randomized clinical trial. JAMA 2016; 316:509–518
12. Hammond DA, Ficek OA, Painter JT, et al. Prospective open-label trial of early concomitant vasopressin and norepinephrine therapy versus initial norepinephrine monotherapy in septic shock
. Pharmacotherapy 2018; 38:531–538
13. Permpikul C, Tongyoo S, Viarasilpa T, et al. Early use of norepinephrine in septic shock resuscitation
(CENSER). A randomized trial. Am J Respir Crit Care Med 2019; 199:1097–1105
14. Coopersmith CM, De Backer D, Deutschman CS, et al. Surviving Sepsis
Campaign: Research priorities for sepsis
and septic shock
. Crit Care Med 2018; 46:1334–1356
15. Vincent JL, Sakr Y, Sprung CL, et al. Sepsis
in European intensive care units: Results of the SOAP study. Crit Care Med 2006; 34:344–353
16. Boyd JH, Forbes J, Nakada TA, et al. Fluid resuscitation
in septic shock
: A positive fluid
balance and elevated central venous pressure are associated with increased mortality
. Crit Care Med 2011; 39:259–265
17. Teixeira C, Garzotto F, Piccinni P, et al.; NEFROlogia e Cura INTensiva (NEFROINT) investigators: Fluid
balance and urine volume are independent predictors of mortality
in acute kidney injury. Crit Care 2013; 17:R14
18. Sadaka F, Juarez M, Naydenov S, et al. Fluid resuscitation
in septic shock
: The effect of increasing fluid
balance on mortality
. J Intensive Care Med 2014; 29:213–217
19. Kelm DJ, Perrin JT, Cartin-Ceba R, et al. Fluid
overload in patients with severe sepsis
and septic shock
treated with early goal-directed therapy is associated with increased acute need for fluid
-related medical interventions and hospital death. Shock
20. Wang N, Jiang L, Zhu B, et al.; Beijing Acute Kidney Injury Trial (BAKIT) Workgroup: Fluid
balance and mortality
in critically ill patients with acute kidney injury: A multicenter prospective epidemiological study. Crit Care 2015; 19:371
21. Sakr Y, Rubatto Birri PN, Kotfis K, et al. Higher fluid
balance increases the risk of death from sepsis
: Results from a large international audit. Crit Care Med 2017; 45:386–394
22. Malbrain MLNG, Van Regenmortel N, Saugel B, et al. Principles of fluid
management and stewardship in septic shock
: It is time to consider the four D’s and the four phases of fluid
therapy. Ann Intensive Care 2018; 8:66
23. Scheeren TWL, Bakker J, De Backer D, et al. Current use of vasopressors in septic shock
. Ann Intensive Care 2019; 9:20
24. Dünser MW, Ruokonen E, Pettilä V, et al. Association of arterial blood pressure and vasopressor
load with septic shock mortality
: A post hoc analysis of a multicenter trial. Crit Care 2009; 13:R181
25. Brown SM, Lanspa MJ, Jones JP, et al. Survival after shock
requiring high-dose vasopressor
therapy. Chest 2013; 143:664–671
26. Lesur O, Delile E, Asfar P, et al. Hemodynamic support in the early phase of septic shock
: A review of challenges and unanswered questions. Ann Intensive Care 2018; 8:102
27. Dalimonte MA, DeGrado JR, Anger KE. Vasoactive agents for adult septic shock
: An update and review. J Pharm Pract 2019
28. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009; 42:377–381
29. Gutsche JT, Mikkelsen ME, McCarthy FH, et al. Veno-venous extracorporeal life support in hemodynamically unstable patients with ARDS. Anesth Analg 2017; 124:846–848
30. Norton EC, Dowd BE, Maciejewski ML. Marginal effects-quantifying the effect of changes in risk factors in logistic regression models. JAMA 2019; 321:1304–1305
31. Lederer DJ, Bell SC, Branson RD, et al. Control of confounding and reporting of results in causal inference studies. Guidance for authors from editors of respiratory, sleep, and critical care journals. Ann Am Thorac Soc 2019; 16:22–28
32. Agresti A, Hartzel J. Strategies for comparing treatments on a binary response with multi-centre data. Stat Med 2000; 19:1115–1139
33. Kahan BC. Accounting for centre-effects in multicentre trials with a binary outcome - when, why, and how? BMC Med Res Methodol 2014; 14:20
34. Khanna A, English SW, Wang XS, et al. Angiotensin II for the treatment of vasodilatory shock
. N Engl J Med 2017; 377:419–430
35. Self WH, Semler MW, Bellomo R, et al. Liberal versus restrictive intravenous fluid
therapy for early septic shock
: Rationale for a randomized trial. Ann Emerg Med 2018; 72:457–466
36. Thomas D, Stram D, Dwyer J. Exposure measurement error: Influence on exposure-disease. Relationships and methods of correction. Annu Rev Public Health 1993; 14:69–93