• 
• 

Three Part Question

In [adult patients with penetrating trauma who are in hemorrhagic shock], does [permissive hypotension resuscitation] decrease [morbidity and mortality]?

Clinical Scenario

A 25 year old man is brought to the Emergency department following a gunshot wound to the abdomen. His BP is 70/50mmHg. He is tachycardic with cool extremities. You wonder about recent studies on permissive hypotension in hemorrhagic shock management. You wonder whether you should aggressively resuscitate the patient with crystalloid or restrict the fluid resuscitation before the patient is taken to the operating room.

Search Strategy

Ovid MEDLINE(R) 1946 to 2013 week 32
Embase 1980 to 2013 Week 32
Pubmed
CINAHL

Medline: [(exp Wounds, Penetrating/ OR exp Wounds, gunshot/ OR exp Wounds, Stab/ OR exp Thoracic injuries/ OR exp Abdominal Injuries/ OR exp Shock, Hemorrhagic/) AND (exp Hypotension, Controlled/ OR permissive hypotension.mp. OR delayed resuscitation.mp. OR hypotensive resuscitation.mp. OR deliberate hypotension.mp. OR low volume resuscitation.mp.)] LIMIT to human and English.
EMBase: [(exp penetrating trauma/ OR exp stab wound/ OR exp gunshot injury/ OR exp thorax injury/ OR exp abdominal injury/ OR exp hemorrhagic shock/) AND (permissive hypotension.mp. OR controlled hypotension.mp. OR delayed resuscitation.mp. OR hypotensive resuscitation.mp. OR deliberate hypotension.mp. OR low volume resuscitation.mp.)] LIMIT to human and English.
CINAHL: (penetrating AND injuries OR gunshot AND wound OR stab AND wound OR thoracic AND injuries OR abdominal AND injuries OR hemorrhagic AND shock) AND (permissive AND hypotension OR controlled AND hypotension OR delayed AND resuscitation OR low AND volume AND resuscitation OR hypotensive AND resuscitation OR deliberate AND hypotension) LIMIT to human and English
Pubmed: ((penetrating injuries OR gunshot wounds OR stab wounds OR thoracic injuries OR abdominal injuries OR hemorrhagic shock)) AND (permissive hypotension OR delayed resuscitation OR controlled hypotension OR hypotensive resuscitation OR deliberate hypotension OR low volume resuscitation) AND (Humans[Mesh] AND English[lang])

Search Outcome

Altogether 327 papers found in Pubmed, 72 papers found in Medline, 28 papers found in CINAHL, 73 papers found in Embase. The titles of the search results were reviewed and the relevant abstracts were further reviewed. 5 papers were found of relevance to this clinical question, one of which (Martin et al.) is a preliminary report of another (Bickell et al.), therefore only 4 papers are presented below.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Bickell WH, Wall MJ, Pepe PE, Martin RR, Ginger VF, Allen MK, Mattox KL. 1994 US	598 patients (age≥16) with penetrating torso injuries (gunshot or stab wound) who presented with a pre-hospital systolic blood pressure ≤ 90mm Hg. 289 received delayed fluid resuscitation (no fluid until taken to operating room). 309 received immediate fluid resuscitation. Exclusion criteria: patients with a revised trauma score of zero at the scene of injury, fatal gunshot wound to the head, and patients with minor injuries not requiring operative intervention.	PRCT, level 1b	Survival to discharge	70% in delayed resuscitation group vs. 62% in immediate-resuscitation group, P=0.04	Not blinded. Randomization by alternating day of month. Formal randomization method was not used. 22 patients from delayed fluid group given fluid in violation of study protocol included in intention to treat analysis.
			Patients with ≥1 postoperative complications (ARDS, sepsis, ARF, coagulopathy, wound infection, pneumonia)	23(18-29)% in delayed resuscitation group vs. 30(25-36)% in immediate-resuscitation group, P=0.08
Morrison CA, Carrick MM, Norman MA, Scott BG, Welsh FJ, Tsai P, Liscum KR, Wall MJ, Mattox KL. 2011 US	90 patients in hemorrhagic shock (SBP ≤90 mmHg) who received same standard workup and care in ED were randomized at OR door for intraoperative resuscitation of either low mean arterial pressure, LMAP, (MAP=50mmHg) or high mean arterial pressure, HMAP, (MAP=65mmHg). Exclusion criteria: Age>45 or <15; pregnant women; incarcerated individuals; known history of previus MI, CAD, renal or cerebral vascular disease; unable to definitively rule out brain injury based on mechanism of injury, clinical exam and/or CT scan of the head; patient who wears opt out bracelet; patient's legal representative does not consent to participation.	PRCT (preliminary result only), level 2b	Overall mortality at 30 d	23%(10/44) in LMAP vs. 28%(13/46) in HMAP (p>0.05)	Preliminary data of a PRCT. small sample size difficult to assess statistical significance. Protocol inhibits pharmacologically lowering MAP to target if patient main a higher blood pressure on his or her own. Mechanism of injury differ significantly between two arms. Severity of injury difficult to assess and can be biased. This paper focuses on intraoperative fluid resuscitation rather than fluid resuscitation during prehospital period or in the Emergency Department.
			Post operative coagulopathy	60.5%(23/38) in LMAP vs. 61.1%(22/36) in HMAP (p=0.93)
			Post operative thrombocytopenia	39.5%(15/38) in LMAP vs. 22.2%(8/36) in HMAP (p=0.09)
			Post operative anemia	42.1%(16/38) in LMAP vs. 47.2%(17/36) in HMAP (p=0.97)
Dutton, R. P., Mackenzie, C. F., & Scalea, T. M. 2002 USA	110 patients presenting in hemorrhagic shock (SBP<90 mmHg) were randomized to one of two fluid resuscitation protocols: target SBP > 100 mm Hg (conventional) or target SBP of 70 mm Hg (low). There were 55 patients in each group. Fluid therapy was titrated to this endpoint until definitive hemostasis was achieved. Patients were excluded if they were pregannat, had a CNS injury impairing their level of consciousness or motor function, were older than 55, or had a previous medical history of diabetes or coronary artery disease.	PRCT, evidence level 2b	Average SBP during bleeding (mm Hg)	114 ± 12 in conventional SBP group vs. 100 ±17 in low SBP group (p<0.001)	Proposed methodology and was not achieved as the actual SBP was higher than the target SBP in both groups (however, SBP differ significantly between the two groups). Sample size is insufficiently powered to demonstrate statistical differences. Though clinicians were blinded till the point of randomization, the initial decision to include patients may be biased towards including patients with better prognosis which might explain the lack of significant difference in outcomes. Result may be systematically skewed by Hawthorne effect wherein both groups received more attentive care than usual (more cautious crystalloid infusion due to enrolment to the study). Patient demographics (including age, type of trauma, sex) is heterogeneous between the conventional and low SBP groups. Furthermore, prognostic index (ie. ISS score) prior to intervention differs significantly between the two groups which suggest ineffective randomization. Primary endpoint (ie. in hospital mortality) is direct and relevant but may be too broad to discriminate subtle differences in outcome between groups. Additional surrogate markers could be helpful. This study included both blunt trauma and penetrating trauma and did not specific the site of injury. Thus it doesn’t specifically answer our clinical question which specifies penetrating trauma of the torso.
			Length of active haemorrhage (h)	2.97 ± 1.75 in conventional SBP group vs. 2.57 ± 1.46 in low SBP group (p=0.20)
			Number of death	4 in each group
			Average injury severity score (ISS)	19.55 ± 11.6 in conventional SBP group vs. 23.91 ± 13.8 in low SBP group (p=0.08)
			Predictd survival rate (TRISS methodology)	94.0 ± 12% in conventional SBP group vs. 90.2 ± 17% in low SBP group (p = 0.18)
			Actual survival rate (%)	92.7% in both groups
Ley, E. J., Clond, M. A., Srour, M. K., Barnajian, M., Mirocha, J., Margulies, D. R., and Salim, A. 2011	3137 patients who received crystalloid resuscitation in the ED were categorized into elderly (age, 70-99 years) and non-elderly groups (age, 20-69 years) based on age. There were 2866 patients in the nonelderly group and 271 patients in the elderly. Both groups were analyzed using multivariate logistic regression to investigate the relationship between mortality and volume replacement (using fluid cutoff thresholds of 1L, 1.5L, 2L, and 3L). Gender, age, injury severity score (ISS), Glasgow coma scale (GCS), admission systolic blood pressure (SBP), and fluid replacement volume were considered for inclusion in the logistic regression analysis. Patients who were dead on arrival or with any data missing (gender, age, ISS, GCS, SBP, or IV volume) were excluded from the analysis.	Retrospective study, level of evidence 2b	Mortality odds ratio for IVF≥1L compared to IVF <1L in nonelderly group	1.69 (95% CI, 1.00-2.87), p=0.051	Confounders including injury severity, GCS, SBP, and age are controlled using multivariate regression analysis, which is a less rigorous method than the use of a randomized controlled trial. Significant confounders other than the ones listed may be present. Small number of subjects included in the elderly group. This study included patients with any type of trauma to any body part (head, face, chest, abdomen, extremities) which doesn’t specifically answer our clinical question which specifies penetrating trauma of the torso.
			Mortality odds ratio for IVF≥1.5L compared to IFV<1.5L in nonelderly group	2.09 (95% CI, 1.31-3.33), p=0.002
			Mortality odds ratio for IVF≥2L compared to IFV<2L in nonelderly group	2.27 (95% CI, 1.41-3.65), p=0.0007
			Mortality odds ratio for IVF≥3L compared to IFV<3L in nonelderly group	2.69 (95% CI, 1.53-4.73), p=0.00006
			Mortality odds ratio for IVF≥1L compared to IVF <1L in elderly group	1.10 (95% CI, 0.48-2.49), p=0.82
			Mortality odds ratio for IVF≥1.5L compared to IFV<1.5L in elderly group	2.89 (95% CI, 1.13-7.41), p=0.027
			Mortality odds ratio for IVF≥2L compared to IFV<2L in elderly group	4.57 (95% CI, 1.55-13.53), p=0.006
			Mortality odds ratio for IVF≥3L compared to IFV<3L in elderly group	8.61 (95% CI, 1.55-47.75), p=0.014

Comment(s)

The traditional approach to resuscitation of hypotensive shock from hemorrhage involves IV bolus of large volume of isotonic crystalloid to restore blood pressure. Contrary to this approach, recent evidence suggest that rapid fluid resuscitation in hemorrhagic shock may actually increase mortality. The mechanism is thought to be that high-volume fluid resuscitation in hemorrhagic shock may increase mortality by increasing intravascular volume thus exacerbating hemorrhage, diluting blood in circulation causing coagulopathy, and ultimately reducing oxygen delivery to tissue (Ley et al., 2011). There has been several studies using animal models that showed hypotensive resuscitation result in decreased hemorrhage and better survival outcome compared to aggressive fluid resuscitation during uncontrolled hemorrhage. Clinical studies of permissive hypotension in the resuscitation of trauma patients with penetrating torso injuries showed similar trends. The RCT by Bickell et al. on penetrating torso injuries showed a statistically significant overall mortality rate that was lower in delayed fluid resuscitation group compared to that of the immediate fluid resuscitation prior to operative control of hemorrhage. The most recent and still ongoing RCT by Morrison et al. focuses on the intraoperative fluid resuscitation instead of prehospital and ED fluid resuscitation. The preliminary result did not show any significant outcome due to limited sample size. The RCT by Dutton et al. suffered from several methological flaws and did not find statistically significant difference in mortality between the conventional SBP group and low SB group likely due to a small sample size. The retrospective study by Ley et al. observed that IV fluid replacement of 1.5L or more in the ED was an independent risk factor for mortality for trauma patients. However, the retrospective study did not specify hemorrhagic shock and penetrating trauma in the inclusion criteria thus less applicable to our clinical question. Furthermore, the methodology to control for confounders is less rigorous than an RCT. A search through clinical guidelines shows that hypotensive resuscitation is "accepted" in a national guideline, clearhouse guideline. http://www.guideline.gov/content.aspx?id=37846&search=permissive+hypotension Trauma injuries can have a number of different mechanisms (penetrating vs. blunt) and locations (brain vs. torso) of. The effect of hypotensive resuscitation may be different depending on different types of trauma. Based on the RCT by Bickell et al., permissive hypotension seemed to be associated with better outcome in penetrating torso trauma but this cannot be extrapolated to other types of injuries. More clinical research on this topic is needed.

Editor Comment

JS

Clinical Bottom Line

Delayed fluid resuscitation prior to surgery is safe and is associated with better survival and shorter hospital stay compared to immediate fluid resuscitation in the setting of hemorrhagic shock from penetrating torso injuries that requires surgical operation. These results should not be extrapolated to all age groups, to pregnant patients, to hypotensive patients with blunt trauma or severe head injuries, or to rural trauma care settings.

Level of Evidence

Level 2 - Studies considered were neither 1 or 3.

References

1. William H. Bickell, Matthew J. Wall, Jr., Paul E. Pepe, R. Russell Martin, Victoria F. Ginger, Mary K. Allen, and Kenneth L. Mattox Immediate versus delayed fluid resuscitation for hypotensive patients with penetrating torso injuries. New England Journal of Medicine 1994; 331(17), 1105-1109
2. Morrison CA, Carrick MM, Norman MA, Scott BG, Welsh FJ, Tsai P, Liscum KR, Wall MJ Jr, Mattox KL. Hypotensive resuscitation strategy reduces transfusion requirements and severe postoperative coagulopathy in trauma patients with hemorrhagic shock:Preliminary results of a randomized controlled trial Journal of Trauma-Injury Infection & Critical Care 2011; 70(3), 652-663.
3. Dutton, R. P., Mackenzie, C. F., & Scalea, T. M. Hypotensive resuscitation during active hemorrhage: impact on in-hospital mortality. Journal of Trauma - Injury, Infection and Critical Care 2002; 52(6), 1141-1146
4. 3. Ley, E. J., Clond, M. A., Srour, M. K., Barnajian, M., Mirocha, J., Margulies, D. R., and Salim, A. Emergency department crystalloid resuscitation of 1.5 L or more is associated with increased mortality in elderly and nonelderly trauma patients Journal of Trauma - Injury, Infection and Critical Care 2011; 70(2), 398-400