Aggressive fluid resuscitation is required for the management of traumatic hemorrhagic shock (THS). Progression of hemorrhagic hypotension to an uncompensated condition results in deleterious complications such as intense inflammatory response and ischemia-reperfusion injury (1–3). However, aggressive administration of fluids may disrupt the formation of thrombi, increase hemorrhage, and reduce patient survival (4). Thus, optimized fluids amount before the definitive treatment of hemorrhage has been discussed in trauma resuscitation strategy (5).
Vasopressor may be used in THS patients to increase blood pressure and reduce the volume of fluid resuscitation. European guidelines, based on limited evidence, recommend the use of vasopressor for THS only in the presence of life-threatening hypotension (6). However, in other clinical guidelines, there is no recommendation for the early use of vasopressor in THS patients (7).
The usefulness of vasopressor for THS has been suggested by limited experimental data (8–10), with only one clinical study using low-dose vasopressin (11). On the other hand, previous investigations reported that the use of vasopressor was associated with excessive risk of mortality (12 , 13).
Therefore, the objective of our study was to evaluate the possible association of vasopressor use with mortality in THS patients.
A retrospective cohort study conducted to evaluate the effect of vasopressor use on mortality among THS patients, using recorded data from the Japan Trauma Data Bank (JTDB). This study was approved by the medical ethics committee of the Gunma University Hospital.
Data were obtained from the JTDB, a nationwide trauma registry established in 2003 by the Japanese Association for the Surgery of Trauma and the Japanese Association for Acute Medicine, to improve and ensure the quality of trauma care in Japan (14). During the study period, a total of 260 hospitals including 95% of tertiary emergency medical centers in Japan participated in the JTDB. The JTDB collects 92 data elements (14) relating to patient and hospital information such as patient demographics, premorbid medical conditions, vital signs, Abbreviated Injury Scale (AIS) score, Injury Severity Score (ISS), in-hospital procedures, as well as in-hospital and in-emergency department (ED) mortality. Although the use of vasopressor during the first 24 hours of care is recorded in the JTDB, important information such as the type and dose of vasopressor and the exact time of vasopressor administration in the first 24 hours is not registered.
The study patient flow chart is shown in Figure 1. Data from adult patients with severe trauma were selected for analysis. Inclusion criteria: presence of systolic hypotension (< 90 mm Hg) on arrival at the ED and the requirement of blood transfusion within the first 24 hours. Exclusion criteria: age less than 16 years old or missing age data, patients with an AIS equals to 6 (i.e., unsurvivable injury) for any region, patients with severe traumatic brain injury (AIS for the head > 3) (i.e., the effect of the increased intracranial pressure on elevated pulse pressure), patients with spinal cord injury (AIS = 5, spinal cord injury with the cervical spine, thoracic spine, or lumbar spine) (i.e., presence of neurogenic shock), cardiopulmonary arrest on prehospital arrival or on ED arrival, cardiopulmonary resuscitation (CPR) (i.e., use of adrenaline during CPR), and missing data regarding the volume of blood transfusion within the first 24 hours or missing data on patient survival.
The primary outcome of this investigation was in-hospital mortality, and the secondary outcome was in-ED mortality.
First, normality for each variable was identified using the Kolmogorov-Smirnov test to report the characteristics of participants. In addition, graphs of each variable were plotted to determine the use of either the mean ± SD or median (interquartile range). Subsequently, continuous variables were expressed as medians (interquartile range). Second, we divided patients into two groups: vasopressor use (vasopressor+) and nonvasopressor use (vasopressor–). Comparisons of continuous variables between the vasopressor+ and vasopressor– groups were performed using the Mann-Whitney U test. Categorical variables were expressed as counts and percentages, and the comparisons of each categorical variable between groups were performed using the chi-square test. Outcomes were evaluated by univariate and multiple logistic regression analyses to assess the independent effect of vasopressor use. We carefully selected confounders on the basis of the assumption that none were directly affected by the use of vasopressor with previous reports and clinical importance for multiple logistic regression models (12 , 13). The covariates included age, gender, year of onset, vital signs at admission to the ED (i.e., Glasgow Coma Scale [GCS] value, systolic blood pressure, heart rate, and respiratory rate), ISS, mechanism of injury, cause of injury, use of prehospital IV fluid, and the volume of transfusion. Multicollinearity was also assessed by a variance inflation factor, and the tolerance value was set at less than 2. The goodness of fit was also assessed by the Hosmer-Lemeshow test. Statistical significance was defined as a two-sided p value of less than 0.05 or assessed using a 95% CI in all statistical analyses. All statistical analyses excluding the propensity score (PS) matching were performed using the IBM SPSS Statistics Version 23.0 (SPSS, Chicago, IL).
We chose PS matching analysis because the use of vasopressor was not randomly assigned. A logistic regression analysis was performed to estimate a PS for the prediction of vasopressor use from the available predictors. As to a PS, confounders were carefully selected on previous reports (12–14), and the clinically important confounders were included to estimate PSs. These variables were same variables of the main multiple logistic regression analysis in addition to premorbid medical conditions leading to coagulopathy (liver cirrhosis, hematologic disorders, and use of anticoagulant or antiplatelet drugs). Hemorrhagic treatment (e.g., transcatheter arterial embolization and operations) were not included in the PS because they could be performed after the initiation of vasopressor administration. Revised Trauma Score (RTS) and Trauma and ISS (TRISS) were also not included in the PS because they included identical components such as vital signs and ISS, which we have already included in the PS. PS matching extracted 1:1 matched pairs of subjects using a caliper with 0.0003 with use/or not of vasopressor based on averaged PS. The absolute standardized difference of variables for the estimation of PS was used to assess the match balance, whereby an absolute standardized difference above 0.1 represents meaningful imbalance. In the PS-matched cohort, an univariate logistic regression analysis was performed to the association between the use of vasopressor and outcomes, in addition, generalized linear mixed model was performed to consider a hospital clustering. PS matching and generalized linear mixed model were performed using R software packages (Version 3.4.0; R Development Core Team; R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, http://www.R-project.org/) for Windows.
Two subgroup analyses were performed to identify the potential benefits and demerits of vasopressor use. First, trauma patients with massive transfusion (MT) (> 20 U within 24 hr) (MT model) were selected. Of note, in Japan 1 U of packed RBCs is approximately 140 mL. Second, trauma patients with probability of survival (Ps) greater than 0.5 (survivable injury [SI]) model were selected. In each model, a logistic regression analysis was performed to assess the independent effect of vasopressor use on outcomes by adjusting the same variables of main analysis. We also performed the same analyses for patients who survived 24 hours to adjusted patient’s condition that led to the vasopressor prescription.
Analysis of Missing Data
We excluded the patients who had missing data regarding the amount of blood transfusion and/or hospital outcome. The high rate of missingness could introduce the biases into this study, and we examined the comparisons of variables between the included patients and the excluded patients (Supplemental Table 1, Supplemental Digital Content 1, http://links.lww.com/CCM/E3).
A total of 236,698 patients were registered in the JTDB from January 1, 2004, to December 31, 2015. Of those, 7,411 patients were included due to the presence of systolic hypotension (< 90 mm Hg) on arrival at the ED and the requirement for blood transfusion within the first 24 hours. Finally, a total of 3,551 patients were analyzed. Of the 3,551 patients, 459 patients were treated with vasopressor (vasopressor+ group), whereas the remaining 3,092 patients did not receive vasopressor (vasopressor– group). Patient characteristics are shown in Table 1 and Supplemental Table 2 (Supplemental Digital Content 2, http://links.lww.com/CCM/E4). The median age of patients was 59 years (39–73 yr), and the majority of patients were male (64%; 2,270/3,551 patients). The most common type of trauma was blunt (85%; 3,002/3,549 patients). Regarding vital signs at admission to the ED, GCS value, and systolic blood pressure were significantly lower in the vasopressor+ group than in the vasopressor– group. Heart and respiratory rates at admission were significantly higher in the vasopressor+ group versus the vasopressor– group. The median RTS was 6.4 (95% CI: 5.4–6.8), and RTS in the vasopressor+ group was significantly lower than that in the vasopressor– group (p < 0.001). Median ISS was 22 (13–34), and ISS in the vasopressor+ group was significantly higher than that in the vasopressor– group (p < 0.001). TRISS in the vasopressor+ group was significantly lower than that in the vasopressor– group (p < 0.001). Regarding premorbid medical conditions, significant difference was observed only in the ratio of stroke (p = 0.049). Administration of IV fluids at the prehospital setting was more frequent in the vasopressor+ group than in the vasopressor– group (p = 0.001). Furthermore, laparotomy and angiography of the pelvis were more frequent in the vasopressor+ group than in the vasopressor– group (p < 0.001). A larger volume of blood transfusion within the first 24 hours was administrated in the vasopressor+ group than the vasopressor– group (p < 0.001).
Table 2 showed comparisons of outcomes with and without the use of vasopressor. In-hospital mortality was 43% (198/459 patients) and 16% (481/3,092 patients) in the vasopressor+ and vasopressor– groups, respectively and ED mortality was 6.8% (31/458 patients) and 2.5% (77/3,080 patients), respectively. The use of vasopressor was associated with in-hospital mortality by a multiple logistic regression analysis (odds ratio [OR], 2.172; 95% CI, 1.666–2.833) and was not associated with ED mortality by a multiple logistic regression analysis (OR, 1.391; 95% CI, 0.802–2.413) (Figs. 2 and 3).
Table 1 also showed the baseline characteristics of PS-matched patients (n = 596). The characteristics of PS-matched patients were finely balanced in terms of absolute standardized mean difference, which was less than 0.1 between the groups. In PS-matched patients, in-hospital mortality was 35% (103/298 patients) and 22% (65/298 patients) in the vasopressor+ and vasopressor– groups, respectively. ED mortality was 3.4% (10/298 patients) and 5.7% (17/297 patients), respectively (Table 2). The use of vasopressor was associated with in-hospital mortality (OR, 2.168; 95% CI, 1.442–3.320) and was not associated with in-ED mortality (OR, 0.676; 95% CI, 0.260–1.719) (Figs. 2 and 3).
As subgroup analyses, the use of vasopressor was related to in-hospital mortality in MT patients (OR, 2.029; 95% CI, 1.414–2.911) and was not associated with in-ED mortality (OR, 1.757; 95% CI, 0.846–3.647) Furthermore, the use of vasopressor was related to in-hospital mortality in patients with SI (OR, 1.959; 95% CI, 1.364–2.814) and was not associated with in-ED mortality (OR, 1.351; 95% CI, 0.516–3.533) (Figs. 2 and 3). In restricted analysis to the patients who survived 24 hours, the use of vasopressor was related to in-hospital mortality (OR, 1.719; 95% CI, 1.137–2.599) (Fig. 2).
Analysis of missing data showed that there were only minor differences between the included 3,551 patients and the excluded 1,386 patients; therefore, we thought these minor differences could not affect the conclusion (Supplemental Table 1, Supplemental Digital Content 1, http://links.lww.com/CCM/E3).
This study was a retrospective cohort study conducted to evaluate the effect of vasopressor use on mortality among THS patients using a nationwide trauma database in Japan. Trauma patients exposed to vasopressor had more severe injury than those did not expose to vasopressor. With robust analyses to adjust for severity, the use of vasopressor was significantly associated with in-hospital mortality.
Comparison With Previous Studies
Previous reports have shown that the use of vasopressor for THS was related to mortality after adjusting for multiple confounders (12 , 13). The findings demonstrated the serious risks of vasopressor use in trauma patients. However, those studies may have not been adequately matched with the adjustment of general trauma confounders such as vital signs and trauma scores (12 , 13). The present study showed similar results using standard logistic regression analysis after adjustment of general trauma confounders, despite a lower OR than those of previous studies. In addition, analyses with PS matching and three subgroups were performed. PS matching has been linked to a lower risk of bias for estimating the effects of treatments, exposures, or interventions when using observational or nonrandomized data (15–17). The characteristics of PS-matched patients were finely balanced, and operations for hemorrhage were also balanced in our study (Supplemental Table 2, Supplemental Digital Content 2, http://links.lww.com/CCM/E4). In addition, the analyses appropriately accounted for clustering of patients within hospitals. It also validated the association between the use of vasopressor and outcomes. On the other hand, the findings likely did not apply to the most severely ill patients; for some of these patients, vasopressors may be life-saving and they may be unmatched
The volume of blood transfusion may have been controlled as an important confounder. The indication bias may have remained after controlling trauma severity and fluid resuscitation. Basically, nonresponder to initial fluid could be considered vasopressor use. Therefore, in subgroup analyses, the MT model was designed for those had supposed to receive sufficient blood transfusion. Patients with MT had similarly higher mortality in the vasopressor+ group than the vasopressor– group. In addition, The SI model (Ps > 0.5) subgroup analysis was designed to exclude too severe injury because the vasopressor+ group included more severe cases than the vasopressor– group. This was a novel finding, revealing that the use of a vasopressor was related to mortality in patients with appropriate resuscitation and treatment.
Possible Explanations and Implications
This study showed that the use of vasopressor was related to mortality after adjustment by robust analyses such as PS matching and three subgroup analyses. Patients in the vasopressor+ group received more blood transfusion products than those in the vasopressor– group, and this difference was confirmed in models excluding the PS matching. The analysis on SI patients showed that the vasopressor+ group received MT approximately twice of the vasopressor– group. Vasopressor may have been used as a volume-sparing resuscitation. However, it may consequently lead to a high demand for blood transfusion in the hemorrhagic state. This may be a possible explanation for the association of vasopressor with mortality in the hemorrhagic state (18).
Previous studies have not discussed the disadvantages of vasopressor use within 24 hours (12 , 13). In this study, the definition of mortality within 24 hours was substituted with mortality in the ED, and the use of vasopressor was not related to mortality in the ED. In Japan, trauma surgeons were not always available even in tertiary care centers and the spending time in ED was longer compared with American and European countries (14). This study suggested that the disadvantages of vasopressor use were not obvious within 24 hours; however, it may have eventually deteriorated the state of THS and caused in-hospital death. This finding was novel, considering that previous studies excluded early deaths and focused exclusively on the outcome of in-hospital mortality.
Collectively, the use of vasopressor was related to an increased risk of mortality in THS patients. The marked reduction in blood pressure may tempt physicians to use vasopressor for decompensated hypotension. These findings show that THS patients exposed to vasopressor may escape from ED death. However, these patients finally required larger fluid resuscitation, potentially leading to in-hospital death. Thus, physicians are advised to achieve definitive hemorrhage control without vasopressor.
The authors acknowledge the following limitations of this observational study. First, detailed information on the dose and type of vasopressor, how to use vasopressor such as bolus and/or continuous and where vasopressor was administrated (ED and operation room, and so on) are not registered in the JTDB. Among vasopressors, vasopressin was reported to improve survival in experimental data (8–10), and only low-dose vasopressin possessed clinical usefulness (11). In addition, the duration of vasopressor use was not available in the JTDB. The limited use of vasopressor in emergency trauma surgery was shown not to be related to mortality (19). Second, other confounders relevant to fluid status were not recorded in this study. Although the volume of blood transfusion was used to adjust potential bias regarding the fluid status, other fluid variable such as crystalloid, colloid, and fresh frozen plasma were not adjusted in this study. Early crystalloid resuscitation was reported to be associated with survival (13). Third, in-hospital mortality and in-ED mortality are considered as inadequate fixed times. Jammer et al (20) recommended that the outcomes of mortality should be reported at least 90 days after surgery.
In conclusions, the use of vasopressor in THS patients was related to mortality, although there may be an indication bias. Further studies are warranted to determine potential benefits of vasopressor administration in THS patients.
We especially thank Haruyasu Fujita of Akagi Hospital for helping us with the statistical analysis. We also thank the Japan Trauma Data Bank and all personnel at the participating institutions who contributed the data.
1. Dalibon N, Schlumberger S, Saada M, et al. Haemodynamic assessment of hypovolaemia under general anaesthesia in pigs submitted to graded haemorrhage and retransfusion. Br J Anaesth 1999; 82:97–103
2. Barroso-Aranda J, Zweifach BW, Mathison JC, et al. Neutrophil activation, tumor necrosis factor, and survival after endotoxic and hemorrhagic shock
. J Cardiovasc Pharmacol 1995; 25(Suppl 2):S23–S29
3. Shenkar R, Coulson WF, Abraham E. Hemorrhage and resuscitation
induce alterations in cytokine expression and the development of acute lung injury. Am J Respir Cell Mol Biol 1994; 10:290–297
4. Bickell WH, Wall MJ Jr, Pepe PE, et al. Immediate versus delayed fluid resuscitation
for hypotensive patients with penetrating torso injuries. N Engl J Med 1994; 331:1105–1109
5. Morrison CA, Carrick MM, Norman MA, et al. Hypotensive resuscitation
strategy reduces transfusion requirements and severe postoperative coagulopathy in trauma patients with hemorrhagic shock
: Preliminary results of a randomized controlled trial. J Trauma 2011; 70:652–663
6. Rossaint R, Bouillon B, Cerny V, et al. The European guideline on management of major bleeding and coagulopathy following trauma: Fourth edition. Crit Care 2016; 20:100
7. Rochwerg B, Hylands M, Møller MH, et al. CCCS-SSAI WikiRecs clinical practice guideline: Vasopressor
blood pressure targets in critically ill adults with hypotension and vasopressor
use in early traumatic shock. Intensive Care Med 2017; 43:1062–1064
8. Voelckel WG, Raedler C, Wenzel V, et al. Arginine vasopressin, but not epinephrine, improves survival in uncontrolled hemorrhagic shock
after liver trauma in pigs. Crit Care Med 2003; 31:1160–1165
9. Liu L, Tian K, Xue M, et al. Small doses of arginine vasopressin in combination with norepinephrine “buy” time for definitive treatment for uncontrolled hemorrhagic shock
in rats. Shock 2013; 40:398–406
10. Gazmuri RJ, Whitehouse K, Whittinghill K, et al. Vasopressin infusion with small-volume fluid resuscitation
during hemorrhagic shock
promotes hemodynamic stability and survival in swine. PLoS One 2015; 10:e0130134
11. Cohn SM, McCarthy J, Stewart RM, et al. Impact of low-dose vasopressin on trauma outcome: Prospective randomized study. World J Surg 2011; 35:430–439
12. Plurad DS, Talving P, Lam L, et al. Early vasopressor
use in critical injury is associated with mortality independent from volume status. J Trauma 2011; 71:565–570; discussion 570–572
13. Sperry JL, Minei JP, Frankel HL, et al. Early use of vasopressors after injury: Caution before constriction. J Trauma 2008; 64:9–14
14. Inoue J, Shiraishi A, Yoshiyuki A, et al. Resuscitative endovascular balloon occlusion of the aorta might be dangerous in patients with severe torso trauma: A propensity score analysis. J Trauma Acute Care Surg 2016; 80:559–566; discussion 566–567
15. Ziff OJ, Lane DA, Samra M, et al. Safety and efficacy of digoxin: Systematic review and meta-analysis of observational and controlled trial data. BMJ 2015; 351:h4451
16. Austin PC. The performance of different propensity score methods for estimating marginal hazard ratios. Stat Med 2013; 32:2837–2849
17. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res 2011; 46:399–424
18. Meier J, Pape A, Loniewska D, et al. Norepinephrine increases tolerance to acute anemia. Crit Care Med 2007; 35:1484–1492
19. Van Haren RM, Thorson CM, Valle EJ, et al. Vasopressor
use during emergency trauma surgery. Am Surg 2014; 80:472–478
20. Jammer I, Wickboldt N, Sander M, et al; European Society of Anaesthesiology (ESA) and the European Society of Intensive Care Medicine (ESICM); European Society of Anaesthesiology; European Society of Intensive Care Medicine: Standards for definitions and use of outcome measures for clinical effectiveness research in perioperative medicine: European Perioperative Clinical Outcome (EPCO) definitions: A statement from the ESA-ESICM joint taskforce on perioperative outcome measures. Eur J Anaesthesiol 2015; 32:88–105
hemorrhagic shock; resuscitation; vasopressor
Supplemental Digital Content
Copyright © by 2018 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.