In 2014, trauma was the third leading cause of overall death in the United States.1 It was estimated that approximately 10% of the total US medical expenditures can be attributed to injury-associated medical care.1 Notably, hemorrhagic shock accounted for 30%–40% of the direct causes of trauma death. Besides controlling bleeding sources, administration of intravenous fluids to hemorrhagic shock patients is important as it replenishes intravascular volume.2
Several dilemmas exist regarding fluid administration in trauma patients. First, establishment of vascular access in the prehospital setting is difficult, especially in trauma patients who are prone to vascular collapse.3–5 Second, isotonic crystalloid fluids, such as the Ringer’s lactate solution, may not restore hemodynamic stability rapidly unless administered in large volumes. However, administration of large volumes of isotonic fluids has been associated with multiple system dysfunctions, including cardiac, lung, gastrointestinal tract, coagulation disturbances, and acid-base imbalances.6 Due to their superior plasma expansion characteristics, colloids have been considered as an alternative for fluid replacement in patients with hemorrhagic shock.7 However, studies have found no survival benefits of administration of colloids over crystalloids in trauma patients.8
Hypertonic saline (HS), with or without dextran, showed promising results in animal studies 30 years ago. In 1980, Velasco et al9 demonstrated that an intravenous infusion of highly concentrated sodium chloride solution (2400 mOsmol/L) to a dog model with severe hemorrhagic shock can restore blood pressure and mesenteric blood flow. Subsequent studies found that administration of HS not only can restore hemodynamic stability but can also reverse hormonal and acid-base abnormalities. In addition, serum sodium and potassium concentrations associated with the HS infusion appeared well tolerated in these animal models.10 Nakayama et al11 then showed in a sheep model that small-volume injection of HS can dramatically improve circulatory function during hemorrhagic shock.
The administration of hypertonic solutions offers several advantages over isotonic solutions in the combat field or prehospital settings. The main advantage is that a smaller volume of hypertonic solutions can be used for fluid administration.12–14 In addition, hypertonic solutions are also bacteriostatic, stable under warm conditions, and are unlikely to freeze.15 Unfortunately, conflicting results were found in the clinical trials that compared hypertonic and isotonic solutions.16–27 For this purpose, we performed a systematic review and meta-analysis to summarize current evidence on the effectiveness of administering either HS or hypertonic saline dextran (HSD) in the treatment of hypovolemia in trauma patients.
We conducted a systematic literature search for hypovolemic shock and hypertonic solutions in Cochrane Central Register of Controlled Trials (CENTRAL), Medline (via PubMed) from inception through June 2014, and Embase from inception through June 2014. The first query, the population query, was composed of the following exploded headings and terms: “hypovolemic shock” OR “shock” OR “Shock, Hemorrhagic” OR “Shock, Surgical” OR “Shock, Traumatic” OR “Shock, Hypovolemic.” The second query, the exposure query, was composed of the following exploded headings and terms: “saline solution, hypertonic” OR “hypertonic saline” OR “Hypertonic Solutions, Saline” OR “Saline Hypertonic Solutions” OR “Solutions, Saline hypertonic” OR “Saline Solutions, Hypertonic” OR “Sodium Chloride Solution, Hypertonic” OR “Hypertonic Saline Solutions” OR “Solutions, Hypertonic Saline.” The third query, the outcome query, was composed of the following exploded headings and terms: ‘‘mortality” OR “survival.” There were no restrictions on language, dates, type of article, language, sex, or age. A similar search strategy and search terms were repeated in Embase. In addition, reference lists of potentially relevant reports and reviews were screened to identify other eligible studies. All titles and abstracts from the search were recorded into a file.
Inclusion and Exclusion Criteria
Two reviewers independently identified articles eligible for in-depth examination by using the following inclusion and exclusion criteria. Clinical studies were considered eligible for inclusion if the study included adult patients (aged >18 years) with hypovolemic shock due to trauma or other causes of hemorrhage. The study had to be a randomized controlled trial that compared the outcome between patients receiving hypertonic solutions resuscitation and patients receiving conventional isotonic solution resuscitation. Hypertonic solutions refer to solutions containing greater than 0.09% NaCl with or without dextran. We required the study to be carried out in the prehospital, emergency department, and intensive care unit (ICU) settings. We also required the study reporting survival status at 30 days as one of the main outcome measures. We excluded studies involving sepsis patients because we aimed to summarize evidence related to the effectiveness and safety of hypertonic solutions as an alternative fluid to replace intravascular volume in patients with hypovolemic shock. Study types of case reports, case series, review articles, editorials, letters, meta-analyses, clinical guidelines, those reviewing databases, animal studies, or in vitro studies were also excluded. Any discrepancies between reviewers on articles meriting inclusion were resolved by a consensus meeting of both lead authors. A summary of the study selection is shown in Figure 1.
Data Extraction and Synthesis
All data concerning authorship, country, year of publication, setting (eg, prehospital, emergency department, ICU, and burn center), study population (hemorrhage or trauma patients), study exposure (hypertonic solution or hypertonic solution with dextran), and study outcomes (overall mortality, mortality in 28 days, survival to discharge, complication, and acute respiratory distress syndrome–free survival) were extracted. We also recorded quality indicators of the study design, as required by the Cochrane bias risk tool (Table 1).
We followed the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines for meta-analysis of observational studies in our data extraction, analyses, and report. The studies were pooled using the random effect model described by DerSimonian and Laird.28 Pooled relative risks (pooled RRs) with 95% confidence intervals (CIs) for mortality and HS-/HSD-related complications were summarized using the Mantel-Haenszel-weighted average method. Statistical heterogeneity across studies was tested using the Cochran Q statistic (P < .05) and quantified with the I2 statistic. The I2 statistic describes the variation in the effect estimate that is attributable to heterogeneity across studies, and I2 of 50% or greater indicate a substantial level of heterogeneity. The presence and effect of publication bias were examined using a combination of Begg and Egger tests. The Begg test examines the presence of association among the effect estimates and their variances. The Egger test plots the regression line between precision of the studies (independent variable) and the standardized effect (dependent variable) and tests the null hypothesis of the zero intercept. Both tests evaluate the funnel plot asymmetry. A meta-regression was performed to explore the sources of heterogeneity and assess the treatment effect within each subgroup. We also tested the interaction between the treatment effect and the various subgroups. The subgroups included resuscitation fluids, resuscitation settings, and patient characteristics. We reported the treatment effect within each subgroup, regardless of the significance of a test of the interaction between treatments and prespecified subgroups because the various effect estimates in specific clinical subgroups may be of interest to different specialty clinicians.
Sample Size Calculation for New Clinical Trials
To assess the strength of evidence for a new study that would be required to change the significance (and thus the conclusion) of the meta-analysis, we took a simulation approach to estimate the effects of inclusion for a future study using the method proposed by Ferreira et al29 and Sutton et al.30 We assumed that the new trial would have 2 arms with a distribution of 30-day mortality consistent with the existing trials in current meta-analysis. The data of a new trial were generated stochastically for a prespecified sample size. The simulated new trial would then be added to the current data for an updated meta-analysis. We calculated the power of the updated meta-analysis to change the conclusion of the current meta-analysis for a given sample size. We performed 2 types of simulation. First, we calculated the power of an updated meta-analysis to change the current conclusion at a significance level of .05, given the different specified sample sizes of a new trial. Second, we assumed the smallest worthwhile clinical effect of hypertonic solutions to be 10% mortality reduction in patients with hemorrhagic shock. By setting the threshold of the upper CI of the updated summary RR to be lower than 0.9, we estimated the power of an updated meta-analysis to reach that conclusion for varying sample sizes of a simulated new trial. We used “metan” and “metapow” command package of Stata 11.0 (StataCorp, College Station, TX) for data synthesis, and a 2-tailed P value of <.05 was considered as statistically significant.
Search Results and Study Characteristics
In this systematic review, we identified 570 studies, of which 216 were from PubMed and 354 were from Embase. Screening, based on title and abstract, identified 46 citations (Figure 1). An additional 26 studies were included from reference lists of the identified articles and from other databases. Sixty of the 72 potentially relevant articles were excluded. Twelve studies met the inclusion criteria as shown in Figure 1. Table 1 shows the main study characteristics. A total of 2932 patients were included in the final analysis. Of the 12 trials, 8 were performed in the United States, 2 trials were conducted in Brazil, and 1 trial was each conducted in the United Kingdom and Australia. Eight trials were in the prehospital setting, 3 trials were in the emergency department setting, and 1 trial was in the ICU setting. All included trials used 7.5% NaCl with dextran (4.2% or 6%) or without dextran for resuscitation in the treatment group. Two trials using very different concentrations of NaCl or dextran were excluded because the differences in concentration of NaCl or dextran may change the blood pressure and affect the outcome. The entire trial conducted by Holcroft et al31 was excluded as it used 3% NaCl as the resuscitation fluid, and patients who were administered with 12% dextran in the trial by Vassar et al21 were also excluded. In terms of choice of dextran concentration, all of the studies added 6% dextran except for 2 studies, which added 4.2% dextran.18,24 In the control group, almost all studies used normal saline or lactated Ringer’s solution. Only Maningas et al27 used Plasma-Lyte A. Table 2 summarizes the study design and the quality of included studies. Seven trials were analyzed by intention-to-treat approach, one by per-protocol approach, and other trials did not report the analysis methods. Most of the studies were judged as low risk of bias.
Quantitative Data Synthesis
We divided included subjects into 2 groups: patients who received HS versus those who received isotonic saline (IS), and patients who received HSD versus those who received IS. Compared with patients treated with conventional solutions, patients treated with HS (RR, 0.96; 95% CI, 0.82–1.12) or patients treated with HSD (RR, 0.92; 95% CI, 0.80–1.06) did not have a significantly different mortality risk (Figure 2). Likewise, there was no significant difference in the risk of complications comparing hypertonic solutions to conventional fluids (Table 3). Subgroup analysis on patients treated in the different settings did not change the conclusions.
We explored whether types of fluids, settings of resuscitation, and patient characteristics could be potential sources of heterogeneity by meta-regression analysis. We found none of these variables had significant impact on mortality compared to the reference groups (Table 4). In addition, we did not observe significant publication bias by the Begg and Egger tests and funnel plots (Supplemental Digital Content 1–3, Table 1, Figure 1A, and Figure 1B, http://links.lww.com/AA/B989, http://links.lww.com/AA/B990, http://links.lww.com/AA/B991).
To determine whether data from existing trials are either conclusive or requiring further additional trials, we performed a statistical simulation. We first simulated a new study based on the t-distribution for treatment and control groups with an associated RR obtained from the current meta-analysis. Then we repeated this simulation 1000 times and estimated the power of the updated meta-analysis including the new study. For power calculation, we were testing the superiority of the hypertonic solutions over the conventional crystalloids at a specified P value or confidence limit. Given the largest existing trial containing 632 patients, we first made a realistic assumption that a new trial would roughly double the size of the current largest study, that is, 650 people in both treatment or control arms. We found adding this new simulated trial to the updated meta-analysis would have a mere statistical power of 16.80% (95% CI, 14.53%–19.26%) to change the current conclusion at a significance level of .05. It requires approximately 30,000 people in each arm to be sufficiently powered (power, 83.20%; 95% CI, 80.74%–85.47%) to change current conclusion. Furthermore, we assume a 10% mortality reduction in the RR scale as the smallest worthwhile clinical effect size. A new trial including 650 participants in each arm would have a mere statistical power of 0% (95% CI, 0.00%–0.37%) to reach a conclusive 10% mortality reduction (upper CI lower than 0.9) in the updated meta-analysis. A new trial including up to 3,000,000 patients in each arm still have a low power to reach such conclusion (power, 20.00%; 95% CI, 17.56%–22.62%).
This review included 12 studies that have compared at least 2 different fluid regimens administered to patients with trauma or hemorrhagic shock. Our study does not reveal a significant difference in survival or complication rates between patients treated with either HS or HSD. HS and HSD also do not appear to improve shock-related complications, such as systemic infection. There was no significant heterogeneity in the analysis contained in this review, with the exception of infectious complications.
Wade et al32 conducted a meta-analysis on trauma patients treated with either HS or HSD in 1997. His results showed no significant survival benefits for HS administration, while patients administered with HSD might be associated with better outcomes (odds ratio, 1.2; 95% CI, 0.94–1.57). Our study, however, showed no survival benefit for administration of either HS or HSD. Regardless of the promising results from animal studies, current evidence shows no significant benefit for hypertonic solution administration. Several possible factors may contribute to the discrepancies between human trials and animal studies: HS may have caused uncontrollable bleeding, perhaps by an excessive expansion of the blood volume or by an excessive increase in blood pressure. This event could also be affected by the size of the injured vessel and the infusion rate of HS.33–36
The safety of administering HS or HSD is of great concern. The potential disadvantages of administering HS/HSD include increased bleeding events, cellular or tissue dehydration, neurologic injury or deficit from hypernatremia, hypokalemia, hyperchloremic acidosis, rebound increased intracranial pressure, precipitation or exacerbation of acute renal failure, anaphylactic reactions associated with dextran, and dextran-related coagulopathy.33,37–39 Among the included studies, transient hypernatremia without significant neurologic sequel was observed in 6 studies.12,18,19,23,25,27 Hyperchloremic acidosis was noted in 1 study.26
The smallest worthwhile beneficial effect of a trauma trial is hard to determine by a benefit-to-harm approach because every life counts. Therefore, we used the efficacy threshold in the largest trauma trial to date (clinical randomization of an antifibrinolytic in significant hemorrhage [CRASH-2] trial) as a reference.40 In the CRASH-2 trial, 20,211 trauma patients were recruited to test whether use of tranexamic acid could result in a 10% reduction in the risk of mortality. Using a 10% mortality as the cutoff of the smallest worthwhile beneficial effect, the current meta-analysis would not be conclusively negative, since the upper limits of the CIs of the summary estimates are at or below 0.90. We therefore estimate the possibility of a new trial to change the conclusion of the current meta-analysis. Results of our simulation showed that a sample size of 60,000 is required to result in a change of the current conclusion at a P value of .05, which is highly unlikely in the foreseeable future. A sample size of 6,000,000 would still not be able to reach a conclusion for the hypertonic solutions to have a smallest worthwhile beneficial effect. The methods we used to assess the power for an updated meta-analysis are based on the observed distribution of the current meta-analysis. Other sample size calculation methods for meta-analysis based on a clinically important difference, such as trial sequential analysis, may also give the estimated required total sample size that would be needed to find significance if future studies were done.41,42
This study has several limitations. First, the randomization process in some studies may not be sufficient to remove the patient severity inequity between the treatment and the control group. In the trial by Vassar et al,22 patients in the HSD group had a higher injury severity score, more patients with a decreased conscious level, and more patients with missing blood pressure data in comparison to patients in the isotonic saline group. In the trial by Bulger et al,17 patients in the HSD group had a higher injury severity score and more patients received massive blood transfusions as compared to patients in the control group. A sufficient sample size and stratified randomization technique may be needed for future trials to avoid imbalance in baseline covariates.42 Second, one study that enrolled a mixed population of trauma patients with or without traumatic brain injury was included in our analysis.19 The mechanism of survival and mortality may be quite different between patients with or without traumatic brain injury. However, only 12% (14/114) of HS-treated group and 11% (13/115) of the control group had isolated head injury, and it should not change the conclusion of this study. Third, we defined hypovolemic shock as systolic blood pressure <100 mm Hg in accordance with the advanced trauma life support (ATLS) guideline, which may underestimate the number of shock patients. Blood pressure is a late manifestation of shock and usually does not change until a blood loss up to 30% to 40% of total blood volume.43 Although a more sensitive measure of shock, such as the shock index or trauma associated severe hemorrhage (TASH) score, may identify shock patients at an earlier stage, they were not used as the inclusion criteria by most clinical trials of hemorrhagic shock.41,44 Fourth, heterogeneity is inevitable in all meta-analysis, as the different trials were performed by different teams, in different places, and with different methods.45 Often, heterogeneity in the studies increases the CI, making interpretation of results difficult. We have taken 3 approaches to minimize heterogeneity and to increase the confidence in our results: (1) we used a random effects model to take into account the heterogeneity of the various studies,28,46 (2) we excluded studies that used very different concentration of NaCl or dextran, and (3) we performed a meta-regression analysis to explore the source of heterogeneity and to assess the treatment effect within each of the subgroups. None of the interactions between the treatment effect and the various subgroups were significant, suggesting a low amount of heterogeneity. Thus, we do not think heterogeneity in the studies was present in our results.
Based on the available data, the use of hypertonic solutions or hypertonic solutions with dextran was not associated with a significant increase in mortality among trauma patients with hemorrhagic shock. Based on the simulation results, an updated meta-analysis that includes a new clinical trial sized of under 30,000 patients is not likely to change the current conclusion. Since most of the studies were not designed to confirm the safety of hypertonic solutions, it is difficult to conclude that hypertonic solutions could be considered as a safe alternative. Further well-designed studies looking at both efficacy and safety are needed before we can conclude that hypertonic solutions should be considered as an alternative.
We thank the staff of the Core Laboratories at the Department of Medical Research in National Taiwan University Hospital for technical support and the Medical Wisdom Consulting Group for technical assistance in statistical analysis.
Name: Meng-Che Wu, MD.
Contribution: This author helped contribute to the writing of the paper, and also agreed with the manuscript's results and conclusions.
Name: Tin-Yun Liao, MD.
Contribution: This author helped collect data/do experiments for the study, write the first draft of the paper, and contribute to the writing of the paper and also agreed with the manuscript’s results and conclusions.
Name: Erica M. Lee, BSPS.
Contribution: This author helped collect data/do experiments for the study and contribute to the writing of the paper and also agreed with the manuscript’s results and conclusions.
Name: Yueh-Sheng Chen, MD.
Contribution: This author helped design the experiments/the study, collect data/do experiments for the study, and write the first draft of the paper and also agreed with the manuscript’s results and conclusions.
Name: Wan-Ting Hsu, MS.
Contribution: This author helped analyze the data, collect the data/do experiments for the study, contribute to the writing of the paper, and also agreed with the manuscript’s results and conclusions.
Name: Meng-tse Gabriel Lee, PhD.
Contribution: This author helped collect the data/do experiments for the study, contribute to the writing of the paper, and also agreed with the manuscript’s results and conclusions.
Name: Po-Yang Tsou, MD, MPH.
Contribution: This author helped contribute to the writing of the paper, and also agreed with the manuscript's results and conclusions.
Name: Shyr-Chyr Chen, MD, MBA.
Contribution: This author helped develop research concept and oversee research and also agreed with the manuscript’s results and conclusions.
Name: Chien-Chang Lee, MD, ScD.
Contribution: This author helped design the experiments/the study, analyze the data, develop research concept, and oversee research and also agreed with the manuscript’s results and conclusions.
This manuscript was handled by: Richard P. Dutton, MD.
2. Mock C. Guidelines for Essential Trauma Care. 2004.Geneva, Switzerland: World Health Organization.
3. Byars DV, Tsuchitani SN, Erwin E, Anglemyer B, Eastman J. Evaluation of success rate and access time for an adult sternal intraosseous device deployed in the prehospital setting. Prehosp Disaster Med. 2011;26:127–129.
4. Hartholt KA, van Lieshout EM, Thies WC, Patka P, Schipper IB. Intraosseous devices: a randomized controlled trial comparing three intraosseous devices. Prehosp Emerg Care. 2010;14:6–13.
5. Dubick MA, Kramer GC. Hypertonic saline dextran (HSD) and intraosseous vascular access for the treatment of haemorrhagic hypotension in the far-forward combat arena. Ann Acad Med Singapore. 1997;26:64–69.
6. Cotton BA, Guy JS, Morris JA Jr, Abumrad NN. The cellular, metabolic, and systemic consequences of aggressive fluid resuscitation strategies. Shock. 2006;26:115–121.
7. Salmon JB, Mythen MG. Pharmacology and physiology of colloids. Blood Rev. 1993;7:114–120.
8. Perel P, Roberts I, Ker K. Colloids versus crystalloids for fluid resuscitation in critically ill patients. Cochrane Database Syst Rev. 2013;2:CD000567.
9. Velasco IT, Pontieri V, Rocha e Silva M Jr, Lopes OU. Hyperosmotic NaCl and severe hemorrhagic shock. Am J Physiol. 1980;239:H664–H673.
10. Smith GJ, Kramer GC, Perron P, Nakayama S, Gunther RA, Holcroft JW. A comparison of several hypertonic solutions for resuscitation of bled sheep. J Surg Res. 1985;39:517–528.
11. Nakayama S, Sibley L, Gunther RA, Holcroft JW, Kramer GC. Small-volume resuscitation with hypertonic saline (2,400 mOsm/liter) during hemorrhagic shock. Circ Shock. 1984;13:149–159.
12. Bulger EM, May S, Kerby JD, et al.; ROC Investigators. Out-of-hospital hypertonic resuscitation after traumatic hypovolemic shock: a randomized, placebo controlled trial. Ann Surg. 2011;253:431–441.
13. Kramer GC. Hypertonic resuscitation: physiologic mechanisms and recommendations for trauma care. J Trauma. 2003;54:S89–S99.
14. Dubick MA, Atkins JL. Small-volume fluid resuscitation for the far-forward combat environment: current concepts. J Trauma. 2003;54:S43–S45.
15. Holcroft JW. The hypertonic saline trial: a possible downside to the gold standard of double blinding. Ann Surg. 2011;253:442–443.
16. Bulger EM, May S, Brasel KJ, et al.; ROC Investigators. Out-of-hospital hypertonic resuscitation following severe traumatic brain injury: a randomized controlled trial. JAMA. 2010;304:1455–1464.
17. Bulger EM, Jurkovich GJ, Nathens AB, et al. Hypertonic resuscitation of hypovolemic shock after blunt trauma: a randomized controlled trial. Arch Surg. 2008;143:139–148.
18. Alpar EK, Killampalli VV. Effects of hypertonic dextran in hypovolaemic shock: a prospective clinical trial. Injury. 2004;35:500–506.
19. Cooper DJ, Myles PS, McDermott FT, et al.; HTS Study Investigators. Prehospital hypertonic saline resuscitation of patients with hypotension and severe traumatic brain injury: a randomized controlled trial. JAMA. 2004;291:1350–1357.
20. Younes RN, Aun F, Ching CT, et al. Prognostic factors to predict outcome following the administration of hypertonic/hyperoncotic solution in hypovolemic patients. Shock. 1997;7:79–83.
21. Vassar MJ, Fischer RP, O’Brien PE, et al. A multicenter trial for resuscitation of injured patients with 7.5% sodium chloride. The effect of added dextran 70. The Multicenter Group for the Study of Hypertonic Saline in Trauma Patients. Arch Surg. 1993;128:1003–1011.
22. Vassar MJ, Perry CA, Holcroft JW. Prehospital resuscitation of hypotensive trauma patients with 7.5% NaCl versus 7.5% NaCl with added dextran: a controlled trial. J Trauma. 1993;34:622–632.
23. Younes RN, Aun F, Accioly CQ, Casale LP, Szajnbok I, Birolini D. Hypertonic solutions in the treatment of hypovolemic shock: a prospective, randomized study in patients admitted to the emergency room. Surgery. 1992;111:380–385.
24. Vassar MJ, Perry CA, Gannaway WL, Holcroft JW. 7.5% sodium chloride/dextran for resuscitation of trauma patients undergoing helicopter transport. Arch Surg. 1991;126:1065–1072.
25. Mattox KL, Maningas PA, Moore EE, et al. Prehospital hypertonic saline/dextran infusion for post-traumatic hypotension. The U.S.A. Multicenter Trial. Ann Surg. 1991;213:482–491.
26. Vassar MJ, Perry CA, Holcroft JW. Analysis of potential risks associated with 7.5% sodium chloride resuscitation of traumatic shock. Arch Surg. 1990;125:1309–1315.
27. Maningas PA, Mattox KL, Pepe PE, Jones RL, Feliciano DV, Burch JM. Hypertonic saline-dextran solutions for the prehospital management of traumatic hypotension. Am J Surg. 1989;157:528–533.
28. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–188.
29. Ferreira ML, Herbert RD, Crowther MJ, Verhagen A, Sutton AJ. When is a further clinical trial justified? BMJ. 2012;345:e5913.
30. Sutton AJ, Cooper NJ, Jones DR, Lambert PC, Thompson JR, Abrams KR. Evidence-based sample size calculations based upon updated meta-analysis. Stat Med. 2007;26:2479–2500.
31. Holcroft JW, Vassar MJ, Turner JE, Derlet RW, Kramer GC. 3% NaCl and 7.5% NaCl/dextran 70 in the resuscitation of severely injured patients. Ann Surg. 1987;206:279–288.
32. Wade CE, Kramer GC, Grady JJ, Fabian TC, Younes RN. Efficacy of hypertonic 7.5% saline and 6% dextran-70 in treating trauma: a meta-analysis of controlled clinical studies. Surgery. 1997;122:609–616.
33. Dubick MA, Bruttig SP, Wade CE. Issues of concern regarding the use of hypertonic/hyperoncotic fluid resuscitation of hemorrhagic hypotension. Shock. 2006;25:321–328.
34. Bruttig SP, O’Benar JD, Wade CE, Dubick MA. Benefit of slow infusion of hypertonic saline/dextran in swine with uncontrolled aortotomy hemorrhage. Shock. 2005;24:92–96.
35. Brauer LP, Svensén CH, Hahn RG, Kilicturgay S, Kramer GC, Prough DS. Influence of rate and volume of infusion on the kinetics of 0.9% saline and 7.5% saline/6.0% dextran 70 in sheep. Anesth Analg. 2002;95:1547–1556.
36. Riddez L, Drobin D, Sjöstrand F, Svensén C, Hahn RG. Lower dose of hypertonic saline dextran reduces the risk of lethal rebleeding in uncontrolled hemorrhage. Shock. 2002;17:377–382.
37. White H, Cook D, Venkatesh B. The use of hypertonic saline for treating intracranial hypertension after traumatic brain injury. Anesth Analg. 2006;102:1836–1846.
38. de Jonge E, Levi M. Effects of different plasma substitutes on blood coagulation: a comparative review. Crit Care Med. 2001;29:1261–1267.
39. Moon PF, Kramer GC. Hypertonic saline-dextran resuscitation from hemorrhagic shock induces transient mixed acidosis. Crit Care Med. 1995;23:323–331.
40. Shakur H, Roberts I, Bautista R, et al. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet (London, England). 2010;376:23–32.
41. Rady MY, Smithline HA, Blake H, Nowak R, Rivers E. A comparison of the shock index and conventional vital signs to identify acute, critical illness in the emergency department. Ann Emerg Med. 1994;24:685–690.
42. Roberts C, Torgerson DJ. Baseline imbalance in randomised controlled trials. BMJ. 1999;319:185.
43. Kortbeek JB, Al Turki SA, Ali J, et al. Advanced trauma life support, 8th edition, the evidence for change. J Trauma. 2008;64:1638–1650.
44. Yücel N, Lefering R, Maegele M, et al.; Polytrauma Study Group of the German Trauma Society. Trauma Associated Severe Hemorrhage (TASH)-score: probability of mass transfusion as surrogate for life threatening hemorrhage after multiple trauma. J Trauma. 2006;60:1228–1236.
45. Higgins JP. Commentary: heterogeneity in meta-analysis should be expected and appropriately quantified. Int J Epidemiol. 2008;37:1158–1160.
46. Colditz GA, Burdick E, Mosteller F. Heterogeneity in meta-analysis of data from epidemiologic studies: a commentary. Am J Epidemiol. 1995;142:371–382.