U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Clinical Guideline Centre (UK). Major Trauma: Assessment and Initial Management. London: National Institute for Health and Care Excellence (NICE); 2016 Feb. (NICE Guideline, No. 39.)

Cover of Major Trauma: Assessment and Initial Management

Major Trauma: Assessment and Initial Management.

Show details

10Assessment and management of haemorrhage

10.1. Control of external haemorrhage

10.1.1. Use of haemostatic dressings

Introduction

Uncontrolled haemorrhage is one of the leading causes of death after injury. In the pre-hospital setting it is important to have effective interventions to control haemorrhage before definitive treatment in hospital. Haemostatic dressings are novel treatments developed for the military setting that are now being considered for civilian use. They can be broadly categorised into three classes based on their method of haemorrhage control. Factor concentrators promote clotting through the rapid absorption of the water content of blood. Mucoadhesives act by adhering to tissue and physically sealing bleeding wounds. Procoagulants either activate the clotting cascade or provide clotting factors, such as fibrinogen and thrombin, to the wound site.

Review question: Are haemostatic dressings clinically and cost effective in improving outcomes in patients with haemorrhage in major trauma?

For full details see review protocol in Appendix C.

Table 40PICO characteristics of review question

PopulationChildren and adults with haemorrhage after experiencing a traumatic incident.
Intervention(s)Haemostatic dressings
Comparison(s)A comparison of the above
Standard dressings (with no active ingredients)
OutcomesCritical:
  • Mortality at 24 hours, 30 days/1month and 12 months
  • Health-related quality of life
  • Adverse effects
    • skin burns
    • delayed wound healing
    • necrosis
    • surgical complications
  • Length of ICU stay
  • Blood product use
Important:
  • Patient reported outcomes (psychological wellbeing)
Study designRCTs or systematic reviews of RCTs; cohort/case control studies if RCT evidence is insufficient.

Only cohort/case control studies accounting for important confounding factors will be considered (severity of shock, severity of injury, degree of head injury, age)

Clinical evidence

No relevant clinical studies were identified.

Economic evidence

Published literature

No relevant economic evaluations were identified.

See also the economic article selection flow chart in Appendix E.

Unit costs

A range of the haemostatic agents available are presented below for illustration of costs.

Table 41. Intervention costs.

Table 41

Intervention costs.

Evidence statements

Clinical

No relevant clinical studies were identified.

Economic

No relevant economic evaluations were identified.

Recommendations and link to evidence

Recommendations
18.

Use simple dressings with direct pressure to control external haemorrhage.

Relative values of different outcomesThe critical outcomes for decision making were mortality health-related quality of life, skin burns, delayed wound healing, necrosis, surgical complications, length of ICU stay and blood product use. The GDG considered patient-reported outcomes such as psychological wellbeing to be important outcomes.
Trade-off between clinical benefits and harmsNo clinical evidence was found to evaluate the trade-off between clinical benefits and harms of haemostatic dressings for control of haemorrhage.

The potential benefits of haemostatic dressing are rapid control of bleeding and therefore, better patient outcomes, although, side effects, have been reported.
Trade-off between net health benefits and resource useNo economic evidence was identified for this question.

Haemostatic dressings are designed as pre-hospital tools to control bleeding in situations where there is non-compressible bleeding. They can be classified into 3 groups by mechanism of action: factor concentrators, mucoadhesive agents and procoagulant supplementors. Haemostatic dressings are more expensive than standard dressings and can vary from £30 to £160 depending on the type.

Some types can have side effects.. Poor control of bleeding could also lead to longer hospital/ICU stay and more use of blood components. Any additional benefit has to be weighed up against the cost and potential adverse events.

The population that will benefit from haemostatic dressings in a civilian population is likely to be very small, and additionally, the products also have an expiry date meaning it can be costly to replace them even if they have not been used.
They are unlikely to stop uncompressible active arterial bleeding, and therefore, may be used in combination with other interventions to control bleeding, such as tourniquets applied to a limb. However, where these dressings may be useful, if effective, is when the injury is in a location in which it is difficult to apply a tourniquet.

The GDG felt that given the small population that they will be used on and the resource impact in purchasing them (which can be expensive and then replacing them [used or unused]), as well as there being no evidence to suggest benefit above standard dressings and direct pressure – a recommendation to use simple dressings and direct pressure was considered appropriate. This could lead to cost savings.

Haemostatic dressings are even less likely to be used in the paediatric population where penetrating injury to peripheral vasculature is extremely rare.
Quality of evidenceNo relevant clinical studies were identified.
Other considerationsHaemostatic dressings are most commonly used in the military environment where they can be utilised almost immediately after injury. The GDG agreed that a key difference between the civilian and military settings is that the administration of haemostatic dressings will be on average much more delayed in the civilian setting. The GDG considered that this may mean an unproven treatment in the military setting is less effective in the civilian setting. The GDG indicated that currently haemostatic dressings are rarely used by ambulance trusts and there is limited knowledge around their correct use. This, combined with a high cost and limited shelf life, led the GDG to agree that there is no reason to recommend these products. In the absence of any evidence in favour of haemostatic dressings, the GDG did not believe that they offered any improvement over and above standard dressings with direct pressure.

The GDG did not identify any considerations specific to children.

10.1.2. Use of tourniquets in major trauma

Introduction

The utility and safety of using pneumatic or mechanical tourniquets in a civilian trauma situation is widely debated and, at times, controversially supported predominately by anecdotal military evidence. While the preservation of life often takes precedence over the potential expense of a limb, there are many potential complications (morbidity, disability and amputation) resulting from inappropriately applied tourniquets or tourniquets left on for excessive amounts of time, when perhaps manual direct pressure would have sufficed to stem the bleeding. This review attempts to identify evidence on the use of mechanical or pneumatic tourniquets in comparison with each other (to find which may work best) or in comparison with no tourniquet or direct pressure (to find out if their use should be supported at all).

Review question: Is the use of pneumatic or mechanical tourniquets clinically and cost effective in improving outcomes in patients with haemorrhage in major trauma?

For full details see review protocol in Appendix C.

Table 42PICO characteristics of review question

ObjectiveTo determine the optimal type of tourniquet to use in patients with limb trauma haemorrhage in the emergency department.
PopulationChildren, young people and adults who have experienced a traumatic limb injury.
Intervention(s)Pneumatic tourniquets
Mechanical tourniquets
Comparison(s)Each other
No tourniquet/direct pressure
OutcomesCritical
  • Mortality at 24 hours, 30 days/1 month and 12 months
  • Health-related quality of life
  • Blood product use (RBCs, platelets, plasma, cryoprecipitate)
  • Length of ICU stay
  • Adverse effects: amputation, nerve palsies, renal failure.
Important
  • Time to definitive control of haemorrhage
  • Patient reported outcomes (psychological wellbeing)
Study designRCTs or systematic reviews of RCTs; cohort studies that use multivariate analysis to adjust for key confounders (injury severity, age, depth of shock, degree of head injury) or were matched at baseline for these if no RCTs retrieved

Clinical evidence

No relevant clinical studies were identified.

Economic evidence

Published literature

No relevant economic evaluations were identified.

See also the economic article selection flow chart in Appendix E.

Unit costs

Tourniquets can come in the form of mechanical or pneumatic:

  • All mechanical tourniquets are single use. It is possible that they may be re-used if appropriately sterilised, however, given the nature of their use, they often come into contact with blood and are therefore generally disposed of and replaced within clinical practice.
  • Pneumatic tourniquets, although involve a much larger upfront cost, are re-usable, and therefore, the cost per use is likely to be minimal, however, this is dependent upon whether re-usable or disposable cuffs are used, which can be expensive.

Table 43. Tourniquet costs.

Table 43

Tourniquet costs.

Potential adverse events from tourniquets can include amputation of limbs, renal failure and nerve palsies. As further information, the cost of treating these adverse events can be found in Appendix O.

As a brief summary:

Table 44. Summary of adverse event costs.

Table 44

Summary of adverse event costs.

Evidence statements

Clinical

No relevant clinical evidence was identified.

Economic

No relevant economic evaluations were identified.

Recommendations and link to evidence

Recommendations
19.

In patients with major limb trauma use a tourniquet if direct pressure has failed to control life-threatening haemorrhage.

Relative values of different outcomesThe GDG agreed the following critical outcomes to inform decision making: mortality, health-related quality of life, adverse effects, blood product use and length of intensive care stay. Time to definitive haemorrhage control and patient-reported outcomes such as psychological wellbeing were identified as important outcomes.

Mortality was considered the most important outcome. If effective tourniquet use resulted in living with a reduced quality of life due to an adverse effect caused by tourniquet reducing blood flow to a limb (for example, nerve palsies), this was considered preferable to death from uncontrolled bleeding.
Trade-off between clinical benefits and harmsThe GDG discussed the harms and benefits of tourniquets versus standard dressings overall, as well as mechanical versus pneumatic.

Control of catastrophic haemorrhage should be the first stage of assessment and resuscitation of a critically injured person. Whereas, immediate haemorrhage control can be achieved by direct pressure, the decision of when direct pressure should be used over tourniquets was considered controversial as the GDG tried to weigh up the risk and cost of placing a tourniquet on a person who did not require it compared with those that do. If applied incorrectly, tourniquets can result in adverse effects associated with reduced blood flow (amputation, nerve palsies and renal failure) as well as result in increased venous bleeding. The GDG debated whether the harms associated with incorrect application could possibly outweigh the clinical benefit of avoiding mortality and agreed that a reduced quality of life due to an adverse effect was preferable to death.

Hence, to be effective, tourniquets need to be appropriately placed proximal to the wound and applied tightly enough to stop bleeding. Tourniquets in the UK are now available on many UK ambulances, but the GDG did not consider there to be a question about which type of tourniquet should be used in the pre-hospital settings, as they felt it was standard practice to use mechanical tourniquets (improvised by police or bystanders, or commercially available varieties from first response personnel).

It was agreed that when tourniquets are needed, early application would be appropriate and it was acknowledged that ideally they should be readily available close to the incident, for instance in first aid kits and potentially in kit dumps at important strategic sites, such as railway stations and airports.

In hospital departments, pneumatic and mechanical tourniquets are both available. There is no evidence to compare one against the other. On a cost basis, pneumatic tourniquets may be more costly.
Trade-off between net health benefits and resource useNo economic evidence was identified comparing tourniquets with no tourniquet, direct pressure or each other (mechanical versus pneumatic).

Pneumatic tourniquets can range in costs from a few pounds to a few hundred pounds per person, depending on how many people they can be used on over their lifetime, and also depending on whether the cuffs are re-usable or disposable, which can increase the cost dramatically. Mechanical tourniquets cost around £20−£30 for a disposable mechanical tourniquet, some can be re-used which would lower the cost per patient.

Tourniquets can also lead to adverse events which will have associated health and resource implications. A lack of clinical evidence meant that data on the benefit (or lack of) of a tourniquet in controlling bleeding, as well as an estimate of the rate of adverse events, was not available.

The GDG felt that it was more important to be able to save lives even if this meant there would be a risk of adverse events in a small proportion of patients.

The main trade-off around the use of a tourniquet is that an effective tourniquet will control blood loss; however, this must be traded off against the possibility of adverse events from reducing blood flow. These can include amputation of the limb, renal failure and nerve palsies which would cause downstream resource and health implications. The incidence of adverse events from a tourniquet was estimated to be less than 10% by the GDG, with transient nerve palsy being the most reported of the three adverse events. If a tourniquet does not adequately control bleeding then the risk of mortality from this is likely to outweigh the risk of adverse events. As no evidence was identified, there is uncertainty as to the relative benefit of a tourniquet compared with direct pressure, however, use of direct pressure would not be associated with the adverse events associated with a tourniquet, but may not be as effective at controlling blood loss. The timing of application of the tourniquet is also an important factor.

For the different types of tourniquets compared with each other, it is unclear whether the more expensive pneumatic tourniquets would be more effective than the mechanical tourniquets. Additionally no evidence was identified on the difference in adverse events between the two types.

Thus, the cost effectiveness of tourniquets both in general (compared with no tourniquet or direct pressure) and compared with each other remains uncertain. However, the GDG felt mechanical tourniquets were standard practice pre-hospital.

The GDG felt that in practice, tourniquets do provide benefit if used on the right people, however, they tend to be over-used and not necessarily used in the right circumstances, which is having a substantial cost impact to the NHS. The incidence of major trauma's where a tourniquet would be applicable was discussed, as their presence on ambulances have stemmed from their use in war, where the incidence of such injuries requiring a tourniquet would be much larger. It was decided that in order to tailor the use of tourniquets to the applicable population that would benefit most from them, a recommendation suggesting who they should be used on would be helpful for clinicians.
Quality of evidenceNo evidence was retrieved which compared the clinical effectiveness of the different types of tourniquets.
Other considerationsIt was noted in GDG discussion that much of the anecdotal support for tourniquet use is based on military experience. In the military context, tourniquet application takes place immediately or extremely soon after the injury, compared with the civilian context where tourniquet use is comparatively slower due to travel time of the first response personnel. It was suggested that those who would most benefit from tourniquet application in the civilian context usually die from uncontrolled haemorrhage before arrival of emergency personnel. The prevalence of injuries that would benefit from a tourniquet in a civilian setting would also be very small compared with a military setting.

It was noted that current stocking of emergency response vehicles with tourniquets is a disproportionately large allocation of resource in relation to the dearth of evidence for their clinical and cost effectiveness. While this lack of evidence does not suggest that tourniquet use is harmful, it does highlight that spending limited NHS resources on their widespread use, given the small population they might be used on, is questionable.

The GDG did not identify any considerations specific to children.

10.2. Pelvic binders

10.2.1. Introduction

Pelvic fractures is a life-threatening orthopaedic emergency; therefore, achieving pelvic stability is an early and critical goal in order to decrease bleeding, decrease pain, improve mobility and allow for transfers. The overall mortality rate for patients with pelvic fractures is between 10 and 20 percent. That rate jumps to 38 percent if the patient is hypotensive on admission and to 50 percent if the patient has an open pelvic fracture.The pelvic binder is used for the emergency stabilisation of pelvic fractures and haemorrhage control before definitive treatment. Current practice is to give any patient suspected of having a pelvic fracture a pelvic binder at the pre-hospital stage. The suspicion of a pelvic fracture is usually based on the mechanism of injury and so is very non-specific. If a patient has a pelvic fracture they will usually benefit from a pelvic binder, as the binder will stabilise the pelvis, reducing pain and blood loss, which will likely outweigh any adverse effects, such as pressure sores. A pelvic binder is considered to be a safe and non-invasive method of pre-hospital stabilisation that may not cause harm to the individual patient if applied correctly. However over-use will incur the costs of equipment, possible transfer to inappropriate locations or unnecessary investigations with no corresponding benefit in outcome.

10.2.2. Review question: Is the application of pelvic binders pre-hospital in patients suspected of pelvic fracture clinically and cost effective in improving outcomes?

Table 45PICO characteristics of review question

PopulationChildren, young people and adults who are suspected of a pelvic fracture following a traumatic incident.
Intervention(s)Pelvic binders
Comparison(s)No treatment/standard care
OutcomesCritical:
  • Mortality at 24 hours, 30 days/1 month and 12 months
  • Volume of blood components
  • Health-related quality of life
  • Adverse effects (unnecessary imaging)
Important:
  • Patient-reported outcomes (pain/discomfort).
  • Improvement in haemodynamics (blood pressure and heart rate)
Study designRCTs or systematic reviews of RCTs; cohort studies that use multivariate analysis to adjust for key confounders (injury severity, age, depth of shock, degree of head injury) or were matched at baseline for these if no RCTs retrieved.

10.2.3. Clinical evidence

No randomised studies were identified that met the inclusion criteria for our protocol. Two cohort studies were, therefore, included in the review; 47,47 55 these are summarised in Table 46 below. Evidence from these studies is summarised in the clinical evidence summary below (Table 47). See also the study selection flow chart in Appendix D study evidence tables in Appendix G, forest plots in Appendix J, GRADE tables in Appendix I and excluded studies list in Appendix K.

Table 46. Summary of studies included in the review.

Table 46

Summary of studies included in the review.

Table 47. Clinical evidence summary: pelvic binder versus no binder.

Table 47

Clinical evidence summary: pelvic binder versus no binder.

All studies included only participants whose pelvic fracture had already been confirmed with imaging. While Krieg et al 76,76 investigated the use of pre-hospital pelvic binders on this population, the other two studies investigated the use, or otherwise, of pelvic binders while in hospital.

10.2.4. Economic evidence

Published literature

No relevant economic evaluations were identified.

See also the economic article selection flow chart in Appendix E.

Unit costs

Pelvic binders can come in the form of:

  • Improvised pelvic binders (made from a bed sheet for example), which will not have any costs associated with them to the NHS as the materials were already available.
  • Purpose made pelvic binders.

The costs of a sample of purpose made pelvic binders used in practice are provided below to aid the consideration of cost effectiveness. The costs are the unit costs from the supplier of the product. However, each ambulance service can negotiate individual discounts with suppliers.

Table 48. Costs of purpose made pelvic binders.

Table 48

Costs of purpose made pelvic binders.

Also note that these costs are per patient. Although it may be possible to re-use some of the pelvic binders if disinfected appropriately, generally they are treated as single use in practice.

As additional information, cost per unit of packed red blood cells is £122.105

10.2.5. Evidence statements

Clinical

Very low quality evidence from one cohort study comprising 585 participants showed that pelvic binders were clinically effective compared with no binder in terms of reducing mortality, with very serious imprecision.

Very low quality evidence from one cohort study comprising 585 participants showed pelvic binders resulted in a clinically important reduction in total blood product use compared with no binder, with serious imprecision. This effect was greater in the stable patient group.

Very low quality evidence from one cohort study comprising 237 participants demonstrated a clinically important reduction in pRBC use for pelvic binders when compared with no binder, with very serious imprecision.

Economic

No relevant economic evaluations were identified.

10.2.6. Recommendations and link to evidence

Recommendations
20.

If active bleeding is suspected from a pelvic fracture after blunt high-energy trauma:

  • apply a purpose-made pelvic binder or
  • consider an improvised pelvic binder, but only if a purpose-made binder does not fit.
In addition to the major trauma GDG reviewing the clinical and cost effectiveness of pelvic binders, the complex fractures GDG reviewed the accuracy of risk tools for the use of pelvic binders (Complex fractures Clinical guideline see section 7.3) . These recommendations were developed and supported by both of the evidence reviews addressing pelvic binders.

Developing the recommendations
The pelvic binder recommendations were developed across the two guidelines by all members of both GDGs. Evidence reviews were completed for all the guidelines and the separate GDGs reviewed the evidence and drafted recommendations. The overall guideline population of patients with pelvic bleeding meant that similarities and duplication between the draft recommendations were inevitable. The recommendations were taken to the project executive team (PET) for coherence and consistency checking, the PET also had the advantage of identifying gaps in the separate guidelines that had been addressed in another guideline. The PET agreed on a core set of draft recommendations. The core set of recommendations were taken back to each of the separate GDGs for review and agreement. The GDG had access to both evidence reviews.
Relative values of different outcomesThe GDG agreed the following critical outcomes to inform decision making: mortality, health-related quality of life, adverse effects and blood product use. Haemodynamic improvement and patient-reported outcomes, such as psychological wellbeing, were identified as important outcomes.
Trade-off between clinical benefits and harmsTwo retrospective cohort studies were identified. Both studies included only participants whose pelvic fracture had already been confirmed with imaging. One study investigated the use of pelvic binder pre-hospital, and one study investigated the use of pelvic binders in hospital. Two studies reported pelvic binders to be associated with a clinically beneficial reduction in mortality but with very serious imprecision. The evidence also indicated that pelvic binders were associated with a clinically important reduction in blood product transfusion (no imprecision to serious imprecision). One study reported more patients with a pelvic binder requiring massive transfusion compared with no pelvic binder but with very serious imprecision.

The GDG noted that both studies were conducted with patients with confirmed pelvic fractures. As pelvic binders will be applied pre-hospital for patients with a suspected pelvic fracture, the evidence, therefore, only represents a sub-population of patients that will receive a pelvic binder. As a consequence, the GDG felt that pelvic binders should only be applied for patients with strong suspicion of pelvic fracture. However, as the GDG noted that there is currently no accurate method of identifying patients with pelvic fracture pre-hospital, the GDG chose to limit their recommendation to patients with suspected active bleeding following blunt high energy trauma to the pelvis. The GDG felt that this patient group are the patients that are most likely to have a pelvic fracture and most likely to benefit from a pelvic binder. The GDG also noted that this recommendation is supported by the clinical evidence that demonstrated a benefit from pelvic binders in a population with major pelvic fracture and associated haemodynamic instability or shock (except for the patients with stable pelvic fractures included in the Fu study) (that is, not avulsion injuries or fractured pubic rami).

No evidence was identified to evaluate the risk of adverse effects when using pelvic binders. The GDG noted that this may include reduced quality of life due to the binder reducing blood flow to a limb (for example, nerve palsies). However, the GDG felt that the clinical benefit of the pelvic binder in avoiding mortality due to the effective control of bleeding was considered to outweigh the possible harm of adverse effects.

It was noted that, in general, the pelvic binder used should always be proprietary and not improvised due to the risks of adverse events associated with inappropriate force used in the application of improvised ‘sheet’ binders. However, the GDG noted that are pelvic binders available to fit large adults and small children but they may not be available, and therefore, an improvised binder may in these cases be better than no intervention.

The GDG discussed the possibility that, because they are non-invasive and generally perceived as safe, pelvic binders may be applied unnecessarily in some patients with a low index of suspicion for a pelvic fracture as staff choose to ‘err on the side of caution’. The GDG confirmed that the only function of a pelvic binder is to control bleeding. The GDG felt that the over-use of pelvic binders may not cause any harm to the individual patient, but that the NHS would incur the costs of equipment, possible transfer to inappropriate locations or unnecessary investigations with no corresponding benefit in outcome.
Trade-off between net health benefits and resource useNo economic evidence was identified comparing pelvic binders with no treatment/standard care.

Pelvic binders can come in the form of improvised binders (using a bed sheet for example) or purpose-made binders. Improvised binders would have no cost to the NHS whereas purpose-made binders can vary from £30-£80.

The accuracy of the pre-hospital assessment in identifying suspected pelvic fracture will determine how many pelvic binders are used. Over-use of pelvic binders is likely to have an impact on hospital imaging because those who come into hospital wearing a pelvic binder will probably be imaged (this may be negated if the patients are multiply injured and would have been imaged anyway, but our population is not just polytrauma), when actually only a small proportion of them may turn out to have a pelvic fracture. Additionally, a suspected pelvic fracture will affect the triaging decision and the transfer destination as a suspected pelvic fracture can trigger a major trauma call using some current triaging protocols, and result in transfer to a major trauma centre.

A more effective binder is assumed to lead to lower downstream resource use in terms of blood components, adverse events and potentially length of stay.

Two studies were identified comparing pelvic binders with no pelvic binders: Both studies showed that mortality is likely to be lower in the pelvic binder group. The results were somewhat conflicting with regards to resource use as one study (Fu 2013) identified that the pelvic binder group used less blood whereas one study (Ghaemmaghami 2007) identified that the pelvic binder group needed more massive transfusions (more than 6 units of pRBCs in 24 hours, an odds ratio of 1.4). If a pelvic binder does reduce downstream resource use, such as blood components, then it is possible that using less blood components also reduces the risk of transfusion or over-transfusion-related adverse events. It is important to note, however, that both studies were in patients who already had a confirmed pelvic fracture, whereas pre-hospital, not all patients applied a binder will turn out to have a fracture, and it is the benefit of the binder in this entire ‘suspected’ group that has not been identified.

The likelihood of having a clinically relevant pelvic fracture (and therefore really needing to have a pelvic binder) is critical in determining whether routine application of a pelvic binder as a preventative measure is cost effective. If the incidence of pelvic fracture is really low, it is unlikely that the costs involved in routinely applying the binder (plus imaging) will justify the potential health gains or savings of blood components (for which there is conflicting low quality evidence). It was mentioned that open book fractures (which are a type of fracture that would benefit from the use of a binder) have a prevalence of less than 1% in adults, and even less for children.

The GDG agreed that when used in the right population, pelvic binders do have a benefit and particularly purpose-made binders over improvised binders. Currently in practice, pelvic binders are being applied to a large population, most of whom may not benefit and therefore, this is a costly practice. The GDG decided to limit the use of pelvic binders to those patients who are suspected of active bleeding, as this is the group most likely to benefit from a binder.
Quality of evidenceThe quality of the evidence was very low for each outcome reported due to risk of bias, imprecision and indirectness as the study populations were patients with confirmed pelvic fractures. This can be considered an indirect population as the question was focusing on whether the pre-hospital application of binders is clinically and cost effective, and this question remains unanswered because it involves including all the groups this would be applied to; those who do not turn out to have a fracture and have a binder, as well as those who do have a fracture and have a binder. Thus focusing on just the latter group (who are more likely to benefit from the binder) may be over estimating both clinical and cost effectiveness.
Other considerationsThe GDG noted that pre-hospital health professionals and clinicians need to be trained to recognise and monitor for the signs of active bleeding. The importance of dispatching the correct personnel to major trauma incidents was also highlighted.

10.3. Haemostatic agents

10.3.1. Introduction

Uncontrolled haemorrhage is a major cause of death in major trauma. Haemostatic agents prevent, stop or control bleeding and have been shown to improve outcomes in patients following surgery and they may also be effective in major trauma patients.

10.3.2. Review question: Is the use of systemic haemostatic agents clinically and cost effective in improving outcomes in patients with confirmed or suspected haemorrhage in major trauma?

For full details see review protocol in Appendix C.

Table 49PICO characteristics of review question

PopulationChildren, young people and adults who have a suspected haemorrhage following a traumatic incident.
Intervention(s)Factor 7 (recombinant activated factor VII)
Tranexamic acid
Fibrinogen concentrate
Prothrombin complex concentrates
Other anti-fibrinolytic agents
Comparison(s)Nothing
A comparison of the above
In combination
In addition to standard care (Blood components [plasma, RBCs, platelets])
OutcomesCritical:
  • Mortality at 24 hours, 30 days/1 month and 12 months
  • Health-related quality of life
  • Adverse effects
    • venous thromboembolism
    • thrombotic events (myocardial infarction [MI]/stroke, pulmonary embolism)
    • over-transfusion related morbidity
    • infections
  • Blood product use:
    • Red blood cells (RBCs)
    • platelets
    • plasma
    • cryoprecipitate
Important:
  • Time to definitive control of haemorrhage
  • Patient-reported outcomes (psychological wellbeing).
Study designRCTs or systematic reviews of RCTs

10.3.3. Clinical evidence

A relevant Cochrane review118,119 was identified but additional outcomes were specified in this protocol to that reported in the review. The review was checked for included studies. Four studies were included in this review127,127; 16,16; 61,61; 41,42 . Two papers reported on two parallel RCTs each (blunt and penetrating trauma populations) 16,16; 61,61. The included studies are summarised in Table 50 below. Evidence from these studies is summarised in the clinical evidence summary tables below (Table 51 and Table 52). See also the study selection flow chart in Appendix D study evidence tables in Appendix G, forest plots in Appendix J, GRADE tables in Appendix I and excluded studies list in Appendix K.

Table 50. Summary of studies included in the review.

Table 50

Summary of studies included in the review.

Table 51. Clinical evidence summary: Tranexamic acid versus standard care.

Table 51

Clinical evidence summary: Tranexamic acid versus standard care.

Table 52. Clinical evidence summary: Recombinant factor VIIa versus standard care.

Table 52

Clinical evidence summary: Recombinant factor VIIa versus standard care.

Summary of included studies:

Narrative review

There was no statistically significant difference between tranexamic acid and placebo for the median units of blood product transfused (tranexamic acid 3 [IQR2-6] versus placebo 3 [2-6]; p=0.59) 127,127.

In patients with blunt trauma, recombinant factor VIIa reduced 48 hour RBC requirements by 2.6 units compared with the placebo (p=0.02). There was no difference in patients with penetrating trauma (RBC reduction 1.0 unit; p=0.10) 16,16. No significant differences were observed in either trauma population with respect to administration of FFP, platelets or cryoprecipitate 16,16.

10.3.4. Economic evidence

Published literature

Four economic evaluations were identified with the relevant comparison and have been included in this review.93,113,117,122

One cost effectiveness analysis compared tranexamic acid with placebo and was based on the CRASH-2 trial.117,119 Three cost utility analyses compared Factor VIIa with placebo and were based on the Boffard trial.

These are summarised in the economic evidence profiles below (Table 53 and Table 54) and the economic evidence tables in Appendix H. 93,113,122

Table 53. Economic evidence profile: Tranexamic acid versus placebo.

Table 53

Economic evidence profile: Tranexamic acid versus placebo.

Table 54. Economic evidence profile: Factor VIIa versus placebo.

Table 54

Economic evidence profile: Factor VIIa versus placebo.

One economic evaluation relating to this review question was identified but was excluded due to limited applicability and the availability of more applicable evidence.75,75 This is summarised in Appendix L, with reasons for exclusion given.

See also the economic article selection flow chart in Appendix E.

The Pohar study does not report discounted QALYs, and therefore, the ICER reported in the table has been estimated based on the undiscounted QALYs reported which further limits the applicability of the findings. The incomplete reporting may also be seen as a limitation, as it is uncertain where the mean incremental discounted QALY may lie within the reported confidence interval of -1.50 to 2.95. The benefit of this paper in terms of usefulness for decision making is that it is funded by the Canadian Government (as it is a Canadian Health Technology Assessment) and is, therefore, likely to be more impartial compared with the Morris and Rossaint papers whose authors have conflicts of interest.

Unit costs

Table 55. Intervention costs.

Table 55

Intervention costs.

Dosing is dependent on weight and extent of bleeding, thus costs presented above are per unit and may not be representative of the total dose of intervention needed to treat the patient. Doses will also be re-evaluated post coagulation testing.

The success of the haemostatic agents could also be measured by the amount of blood components used. An estimate of these resources involved can be seen below. Again the units used per patient can vary.

Table 56. Blood product costs.

Table 56

Blood product costs.

Note that for children, the costs of FFP and cryoprecipitate are substantially larger due to the Department of Health recommendations for those born after 01.01.96 should use particular types of FFP and cryoprecipitate that have undergone additional reduction procedures to reduce the risk of viruses. Please see chapter 10.4.4 for more detail on this.

10.3.5. Evidence statements

Clinical

Tranexamic acid

High quality evidence from 1 RCT comprising 10,115 participants showed that tranexamic acid was clinically effective compared with placebo in terms of mortality, with no imprecision.

Moderate quality evidence from 1 RCT comprising 10,115 participants showed that no difference in clinical effectiveness between tranexamic acid and placebo in terms of MI/stroke, with serious imprecision.

Low quality evidence from 1 RCT comprising 10,115 participants showed that no difference in clinical harm between tranexamic acid and placebo in terms of pulmonary embolism or deep vein thrombosis, with very serious imprecision.

High quality evidence from 1 RCT comprising 10,115 participants showed that no difference in clinical effectiveness between tranexamic acid and placebo in terms of blood components transfused, with no serious imprecision.

Recombinant factor VIIa

Very low quality evidence from 2 RCTs comprising 819 participants showed that recombinant factor VIIa was clinically effective compared with placebo in terms of mortality, with very serious imprecision.

Low quality evidence from 1 RCT comprising 560 participants showed there was no difference in clinical effectiveness between recombinant factor VIIa and placebo in terms of MI/stroke, with very serious imprecision.

Moderate quality evidence from 1 RCT comprising 474 participants (blunt trauma) showed there was no difference in clinical effectiveness between recombinant factor VIIa and placebo in terms of venous thromboembolism, with serious imprecision.

Moderate quality evidence from 1 RCT comprising 86 participants (penetrating trauma) showed that recombinant factor VIIa was clinically harmful compared with placebo in terms of venous thromboembolism, with serious imprecision.

Low quality evidence from 1 RCT comprising 560 participants showed there was no difference in clinical harm between recombinant factor VIIa and placebo in terms of pulmonary embolism, with very serious imprecision.

Low quality evidence from 1 RCT comprising 837 participants showed there was no difference in clinical harm between recombinant factor VIIa and placebo in terms of thrombotic adverse events, with very serious imprecision.

Moderate quality evidence from 1 RCTs comprising 554 participants showed that recombinant factor VIIa was associated with clinically important reduction compared with placebo in terms of RBCs use, with serious imprecision.

Moderate quality evidence from 1 RCT comprising 554 participants showed that there was no clinical difference between recombinant factor VIIa and placebo in terms of platelets and cryoprecipitate use, with serious imprecision.

High quality evidence from 1 RCT comprising 554 participants showed that recombinant factor VIIa was associated with clinically important reduction compared with placebo in terms of FFP use, with no serious imprecision.

Low quality evidence from 1 RCT comprising 560 participants showed there was no difference in clinical effectiveness between recombinant factor VIIa and placebo in terms of sepsis, with very serious imprecision.

Economic

One cost effectiveness analysis found that tranexamic acid had a cost per life year gained of £42 compared with placebo in bleeding patients. This study was assessed as partially applicable with potentially serious limitations.

One cost utility analysis found that recombinant activated factor VII was cost effective compared with placebo at a threshold of £20,000 (ICER of £18,825 per QALY) in bleeding patients. This study was assessed as directly applicable with potentially serious limitations.

One cost utility analysis found that recombinant activated factor VII was not cost effective compared with placebo at a threshold of £20,000 (ICER of £21,613 per QALY) in bleeding patients. This study was assessed as partially applicable with potentially serious limitations.

One cost utility study found that recombinant activated factor VII was more costly than placebo (£20,342 more per patient) with incremental QALYs calculated but not reported. This study was assessed as partially applicable with potentially serious limitations.

10.3.6. Recommendations and link to evidence

Recommendations
21.

Use intravenous tranexamic acida as soon as possible in patients with major trauma and active or suspected active bleeding.

22.

Do not use intravenous tranexamic acida more than 3 hours after injury in patients with major trauma unless there is evidence of hyperfibrinolysis.

Relative values of different outcomesCritical outcomes for decision making were mortality, health related quality of life, adverse events (venous thromboembolism, thrombotic events [MI/stroke, pulmonary embolism], over-transfusion-related morbidity and infections) and blood product. Important outcomes were time to definitive control of haemorrhage and patient-reported outcomes.

For tranexamic acid there was no clinical evidence for mortality (24 hours, 12 months), health-related quality of life, venous thromboembolism, over transfusion-related morbidity, infections, blood product use (RBCs, platelets, plasma, cryoprecipitate), time to definitive control of haemorrhage and patient-reported outcomes. For recombinant factor VIIa there was no clinical evidence for mortality (24 hours, 12 months), health-related quality of life, over-transfusion-related morbidity, time to definitive control of haemorrhage and patient-reported outcomes
Trade-off between clinical benefits and harmsTranexamic acid resulted in a clinically important reduction in mortality (high quality) compared with placebo. The GDG noted that the control event rate was lower than in the major trauma population that could potentially be treated with tranexamic acid and the mortality benefits of the drug would be higher than that reported in the trial. There was no data on quality of life. There were no clinically important harms reported in the trial. A post-hoc sub-group analysis (not reported here) suggested clinical harm if tranexamic acid is administered after three hours. The GDG noted that empiric administration of tranexamic acid should be avoided if the patient presented more than three hours after injury. However, patients could still benefit from tranexamic acid after three hours if there was diagnostic evidence of continued hyperfibrinolysis.
The GDG did not recommend factor VIIa because, although there was an observed reduction in mortality, the confidence intervals were consistent with both benefit and harm (very low quality). There were clinically more venous thromboembolic events in the patient with blunt trauma. There was a reduction in RBCs (serious imprecision) and FFP associated with recombinant factor VIIa. The GDG felt that the potential for increased thromboembolism reported in non-trauma populations without obvious improvement in survival meant that recombinant factor VIIa could not be recommended.
Trade-off between net health benefits and resource useOne economic evaluation comparing tranexamic acid with standard care, and three economic evaluations comparing recombinant factor VIIa with standard care were identified.

For tranexamic acid, the outcome of the study was cost per life year gained rather than cost per QALY, and as such, cannot be assessed using the £20,000 cost per QALY decision rule. This paper was assessed as partially applicable with potentially serious limitations.

The economic evidence identified showed that particularly for recombinant factor VIIa, cost effectiveness is uncertain, as two studies identified comparing recombinant factor VIIa with placebo generally found that the ICER is just as likely to be under £20,000 as over £20,00093,122 (these papers were assessed as being directly applicable and partially applicable respectively, and both had potentially serious limitations), and one study suggested that it was less likely to be cost effective 113,113 (this paper was assessed as partially applicable with potentially serious limitations).

Tranexamic acid has a substantially lower cost compared with factor VIIa, and the clinical evidence showed a higher risk of adverse events for factor VIIa.

The use of tranexamic acid is common practice; therefore, a recommendation in favour of tranexamic acid is not expected to have a large cost impact.
Quality of evidenceOne RCT was identified comparing tranexamic acid with standard care. The RCT was a very large multicentre trial with low risk of bias. Outcomes were graded from high to low quality. Three studies reporting on two RCTs (2×2 parallel RCTs in the blunt and penetrating trauma populations) were identified comparing recombinant factor VIIa with standard care. One RCT was at high risk of bias (Boffard) and one at low risk of bias (Hauser). Outcomes were graded from high to low quality.

Economic evidence
The health economic evidence was based on the CRASH-2 trial in the case of evaluating TXA 117,119 and the Boffard trial 16,16 in the case of recombinant factor VIIa. As such, the evaluations suffer from the same limitations of the evidence as outlined above. Furthermore, two of the studies on recombinant factor VIIa had conflict of interest 93,122 and no study considered adverse events. Most studies were graded as partially applicable (except Morris 2007 which was directly applicable) and with potentially serious limitations.
Other considerationsThe CRASH-2 trial was in adults, but the GDG felt that the results could be extrapolated to children. Tranexamic acid has been used in non-traumatic paediatric populations with a low incidence of adverse events.

The GDG noted that the evidence on tranexamic acid was in patients with or at risk of significant haemorrhage and emphasised that the importance of pre-hospital personnel recognising the signs of active bleeding or people who at risk of active bleeding.
a

At the time of publication (February 2016), tranexamic acid did not have a UK marketing authorisation for this indication. The prescriber should follow relevant professional guidance, taking full responsibility for the decision. Informed consent should be obtained and documented. See the General Medical Council's Prescribing guidance: prescribing unlicensed medicines for further information.

10.4. Anticoagulation reversal

10.4.1. Introduction

Anticoagulant medicines are most commonly prescribed for people who are at elevated risk of developing blood clots in veins or arteries, and work to prevent this. People who experience a traumatic injury are at increased risk of getting coagulopathy, a condition in which the blood's ability to clot is impaired. When people on pre-existing anticoagulant medication experience a traumatic injury their coagulopathy is exacerbated by the medication and chance of dying increased. Consequently, mortality can be reduced by reversing the effects of any pre-existing anticoagulant medication. There are varying mechanisms by which anticoagulants work and each class of drug requires its own reversal regimen. This clinical question focuses on reversal of four anticoagulant classes: coumarins and phenindione, direct thrombin inhibitors, anti-platelet agents and low molecular weight heparins. Warfarin, the most prescribed anticoagulant, sits within the coumarin and phenindione class.

10.4.2. Review question: What is the most clinically and cost effective regimen for reversal of pre-existing therapeutic anticoagulation (laboratory effect) in major trauma?

For full details see review protocol in Appendix C.

Table 57PICO characteristics of review question

PopulationChildren, young people and adults who have experienced a traumatic incident and who are on pre-existing therapeutic anticoagulation therapy.
Intervention(s)
  • Reversal agents
    • fibrinogen concentrate
    • cryoprecipitate
    • platelets
  • Vitamin K (phytonadione)
  • Fresh frozen plasma (FFP)
  • Prothrombin complex concentrates (PCCs)
  • Recombinant factor VIIa
Comparison(s)To each other
OutcomesCritical:
  • Mortality at 24 hours, 30 days/1month and 12 months
  • Health-related quality of life
  • Adverse events:
    • stroke
    • myocardial infarction (MI)
    • thromboembolism (PA and venous)
  • Reversal of anti-coagulation as measured by laboratory assessment
  • Neurological outcome (brain injured patients)
  • Blood product use
Important:
  • Patient reported outcomes (pain/discomfort, return to normal activities, psychological wellbeing)
Study designRCTs or systematic reviews of RCTs

10.4.3. Clinical evidence

No relevant clinical studies were identified.

10.4.4. Economic evidence

Published literature

One economic evaluation was identified with the relevant comparison and has been included in this review. 57,57 This is summarised in the economic evidence profile below (Table 58) and the economic evidence tables in Appendix H.

Table 58. Economic evidence profile: PCC versus FFP.

Table 58

Economic evidence profile: PCC versus FFP.

See also the economic article selection flow chart in Appendix E.

This study was assessed as partially applicable due to the population not being specific to trauma patients and no discounting being reported. There may be a difference in mortality rates for haemorrhage in a trauma population due to the severity of the haemorrhage, which could change the conclusions of cost-effectiveness. The intracranial bleeding population was felt to be most similar to a trauma haemorrhage population.

A lack of discounting would overestimate the cost of stroke rehabilitation only, since all other costs were assumed to occur in the first year following haemorrhage. However, all QALYs beyond the first year will be overestimated without discounting, which could also change the conclusions about cost-effectiveness. Additionally, as trauma patients are at risk of coagulopathy, this in combination with the anticoagulants increases mortality further, thus the interventions ability to be able to reverse this may have a different level of success, lead to different mortality rates and different resource use to that of the population considered in the paper. Therefore, all these factors could impact the cost effectiveness and reduce the applicability of this paper to the population of this clinical question.

This study has been assessed as having potentially serious limitations due to the lack of evidence for resource use, the small sample sizes for mortality rate estimates and the lack of weighting to calculate the mean mortality rate as well as the assumptions made where evidence is unavailable. Resource use and probabilities of successful reversal were elicited from a group of consultant physicians and so are based on assumptions. There is, therefore, a large amount of uncertainty around these estimates. The uncertainty around mortality rate estimates following PCC treatment has been taken into account by sensitivity analyses; however, the range of values used for gastrointestinal and retroperitoneal haemorrhage, where values were assumed, may not reflect the true uncertainty in mortality. For the mortality estimates following FFP treatment, sensitivity analyses were not performed and there is uncertainty, especially for gastrointestinal haemorrhage, due to the small number of studies found. For the probability distributions in the probabilistic sensitivity analysis, an arbitrary standard deviation of 10% was used for probabilities, which may not reflect the true magnitude of the uncertainty given that these probabilities were based on assumptions. NHS costs were only included in the first year because the authors thought that the costs incurred after this time would be equally likely for both treatments and therefore, the differences would be negligible. However, this does not consider the difference in the number of survivors following treatment, which could cause a greater difference in costs. These limitations could change the conclusion of cost-effectiveness and therefore, must be judged as having potentially serious limitations.

Unit costs

Table 59. Cost of interventions and resources.

Table 59

Cost of interventions and resources.

The costs expressed above are per unit. Dosing can vary depending on, for example; the weight of the patient, if the patient is severely bleeding they may need more products. Therefore, the costs above may not be indicative of the actual volume of a particular product needed.

As well as the costs of the products themselves, additional costs would be involved in the handling and administration of the products, which would apply each time a product is issued, and could vary depending on the type of product (see more in section 10.8.4).

For children, Department of Health recommendations 39 state that children should use particular types of FFP and cryoprecipitate that have undergone additional reduction procedures to reduce the risk of viruses. Any patient born after 01.01.1996 should receive methylene blue (MB)-treated FFP. The MB treatment is a viral inactivation phase and the plasma is sourced abroad. Cryoprecipitate also follows the same rule as FFP for children. An alternative to MB-treated FFP is Octaplas; a solvent detergent treated plasma that undergoes a prion reduction step. Octaplas can be used on adults and children.

MB-treated FFP for children is over 6 times more expensive than standard FPP, and pooled MB-treated cryoprecipitate is over £1000 as the plasma is non-UK sourced, as per the Department of Health recommendation (plasma from outside the UK is known to have lower risk of transfusion transmitted Creutzfeldt Jakob disease [vCJD]). However, the supply of these products is limited due to the difficulty in sourcing sufficient plasma from countries with a lower prevalence of vCJD as they are very few in number and tend not to have available capacity to supply the UK.

The costs of these products can be seen below.

Table 60. Intervention costs for children and young people.

Table 60

Intervention costs for children and young people.

10.4.5. Evidence statements

Clinical

No clinical evidence identified

Economic

One cost-utility analysis found that PCC was cost-effective compared with FFP for emergency warfarin reversal (ICER: £3000 or less per QALY gained for each type of haemorrhage). The paper was assessed as partially applicable with potentially serious limitations.

10.4.6. Recommendations and link to evidence

Recommendations
23.

Rapidly reverse anticoagulation in patients who have major trauma with haemorrhage.

24.

Hospital trusts that admit patients with major trauma should have a protocol for the rapid identification of patients who are taking anticoagulants and the reversal of anticoagulation agents.

25.

Use prothrombin complex concentrate immediately in adults (16 or over) with major trauma who have active bleeding and need emergency reversal of a vitamin K antagonist.

26.

Do not use plasma to reverse a vitamin K antagonist in patients with major trauma.

27.

Consult a haematologist immediately for advice on adults (16 or over) who have active bleeding and need reversal of any anticoagulant agent other than a vitamin K antagonist.

28.

Consult a haematologist immediately for advice on children (under 16s) with major trauma who have active bleeding and may need reversal of any anticoagulant agent.

29.

Do not reverse anticoagulation in patients who do not have active or suspected bleeding.

Relative values of different outcomesThe critical outcomes for decision making were mortality, health-related quality of life, stroke, MI, thromboembolism, reversal of anticoagulation, neurological outcome and blood product use. Important outcomes were patient-reported outcomes, such as pain/discomfort, return to normal activities and psychological wellbeing.
Trade-off between clinical benefits and harmsNo clinical evidence was found to evaluate the trade-off between clinical benefits and harms of reversal regimens for people on pre-existing therapeutic anticoagulant therapy.

The GDG stated that in patients with haemorrhage, effective and immediate reversal of anticoagulant medication is essential. Delays in reversal are associated with an increase in poor outcomes. As such, the GDG agreed it was imperative that anticoagulant reversal is prioritised in actively bleeding patients without necessarily waiting for laboratory results. To ensure this is standard practice in hospitals receiving trauma patients, the GDG considered it important that all hospitals have a policy for the rapid identification and reversal of oral anticoagulant agents.

The GDG recommended PCC because in their opinion it provides rapid effective specific reversal of a vitamin K antagonist compared with other reversal therapies. It is better than plasma because it is comprised of pooled plasma products that have higher levels of coagulation factors and therefore leads to the much more rapid normalisation of INR. PCCs also have the advantage that, in contrast to plasma, they may be held in emergency departments; their volume of infusion is small and not associated with volume-associated sequelae from fluid overload. Furthermore, faster normalisation of INR is possible with PCCs as due to faster preparation (no thawing required) and faster infusion of the product.

The GDG could not recommend PCC for reversal of other causes of anticoagulation in trauma patients as its safety and efficacy is less known.

Reversal of anticoagulation therapy can result in significant adverse effects; including stroke, MI and thromboembolism, and it is important it is not used in patients that are not actively bleeding.
Trade-off between net health benefits and resource useOne economic evaluation was identified comparing PCC with FFP (Guest 2010). The study was a decision tree model capturing the success of reversal of warfarin for three types of haemorrhage (intracranial, gastrointestinal and retroperitoneal), and the probability of requiring an additional warfarin reversal treatment when the initial attempt is unsuccessful. The population was not major trauma patients, instead they were bleeding because of the therapeutic over-anticoagulation (warfarin). The study showed that PCC was cost effective in all three patient groups.

Limitations of the study included a conflict of interest (the study is funded by the manufacturers of PCC), resource use and some mortality based on assumptions, and methodology not always clear. This study was rated as partially applicable with very serious limitations.

Intracranial haemorrhage was felt to be the most similar population to a trauma population, however, the resource use would be significantly different, such as admission longer than 2 days (most likely weeks), and more than 5% would require an operation.

This may have an impact on the conclusions as for those that survive the reversal, the resource use will then be more costly, however, this would be weighed up against the QALYs that are being accrued from those patients still alive. The effect of this on the overall cost effectiveness is therefore uncertain.

Initially, the study was presented to the GDG as having potentially serious/very serious limitations. The group felt that, although the interventions were relevant, it was discussed how the population groups included in the paper were not directly applicable to trauma, with the most applicable being intracranial haemorrhage as discussed above. This along with the limitations meant it was downgraded to very serious limitations. The GDG also agreed that the conclusions of the study were feasible and in line with what they expected.

Costs of the interventions were also presented to the GDG and it was highlighted that notably PCC is expensive, with a typical dose (30 units/kg for a 75 kg person – considered to be a relatively high dose) costing around £1,350. Factor VIIa is also expensive at several hundred pounds, whereas FFP is not as expensive. FFP is a blood product and has to be de-thawed which adds time, and issues of wastage and time spent cross-matching to blood arise (to note; for children, FFP is more expensive because a special type has to be used that goes through a reduction step to reduce viruses). Whereas, PCC can be administered immediately and does not have the timing issues and risks associated with FFP because it does not need to be matched to blood type. Note that, as well as the costs of the products, there will also be administration costs related to handling, and laboratory costs.

Although there is a large population that are on anticoagulants, notably a proportion of the elderly, the number of patients who have had a trauma and are on anticoagulants is likely to be small.

The GDG acknowledged the limitations of the included study and based their recommendation of PCC as opposed to FFP on other considerations such as its rapid effect and fewer complications. Since there is less certainty on its effectiveness in other populations, this intervention was recommended only for people with major trauma who have active bleeding and need emergency reversal of a vitamin K antagonist.

PCC is used in current practice to reverse anticoagulation of warfarin; therefore, this recommendation is not expected to have a cost impact.
Quality of evidenceNo relevant clinical studies were identified.

Economic evidence
The health economic evidence was based on the single paper identified. The population was an indirect population of which intracranial haemorrhage was felt to be the most applicable to a trauma population. The study had many limitations, such as a conflict of interest, many assumptions were made and no adverse events considered. It was rated as partially applicable with very serious limitations.
Other considerationsThe GDG stated that there are currently no clear strategies for reversing novel oral anticoagulants (including apixaban, dabigatran etexilate and rivaroxaban) or direct thrombin inhibitors and these patients should be discussed immediately with a haematologist.

There was no clinical evidence evaluating treatments for reversing anticoagulants in children and the GDG discussed the possibility of duplicating the advice for adults. The GDG agreed that while treatment for children could be the same as for adults because of the very small numbers of involved and the potential to do harm (for example, children with replacement heart valves), it was important that any treatment is discussed immediately with a consultant haematologist.

10.5. Haemorrhage shock prediction/risk tools

10.5.1. Introduction

Haemorrhagic shock is associated with high mortality. The early detection of haemorrhagic shock and patients requiring massive transfusion by the application of a risk tool could substantially improve patient outcomes.

10.5.2. Review question: What is the most accurate risk tool to predict the need for massive transfusion in patients with major trauma (pre-hospital and hospital)?

For full details see review protocol in Appendix C.

Table 61PICO characteristics of review question

PopulationChildren, young people and adults who have experienced a traumatic incident.
Target conditionHaemorrhagic shock
Index test(s)/comparator(s)Pre-hospital and hospital:
  • Clinical risk scores
  • ABC score
  • TASH score
  • PWH score
  • McLaughlin score
  • Emergency transfusion score
  • Shock Index
  • Shock Classification (part of ATLS protocols)
Reference standard(s)Massive transfusion
Statistical measure/outcomesDiagnostic accuracy
Outcomes from false positive/false negative results, adverse effects
Study designObservational studies

10.5.3. Clinical evidence

Nine studies were included in the review18,18; 22,22;26,27; 77,77; 86,86; 90,92; 109,109; 114,114; 137,137 these are summarised in Table 62 below. Evidence from these studies is summarised in the GRADE clinical evidence profile below. See also the study selection flow chart in Appendix D study evidence tables in Appendix G, forest plots in Appendix J, GRADE tables in Appendix I and excluded studies list in Appendix K.

All of the studies were retrospective cohort studies in the adult hospital population. Two of the studies were in military populations22,22; 86,86 with the rest of the studies in the civilian population.

In the Brockamp study18,18 in order to compare the clinical risk tools, the area under the receiver operating characteristic curves (AUCs) were calculated and the cut-off (threshold) with the best relationship between sensitivity and specificity was used to recalculate sensitivity and specificity for each tools.

10.5.3.1. Summary of included studies

Table 62. Summary of studies included in the review.

Table 62

Summary of studies included in the review.

Table 63. Diagnostic evidence profile: Haemorrhagic shock risk prediction.

Table 63

Diagnostic evidence profile: Haemorrhagic shock risk prediction.

10.5.4. Economic evidence

Published literature

No relevant economic evaluations were identified.

See also the economic article selection flow chart in Appendix E.

10.5.5. Evidence statements

Clinical

ABC

When diagnostic meta-analysis was conducted on Very low quality evidence from 4 studies (6 validation data sets) with 3553 participants pooled sensitivity (95% CI) and specificity (95% CI) of the ABC were 0.72 (0.45 to 0.91) and 0.88 (0.76 to 0.95), respectively.

Very low quality evidence from 1 study of 5147 participants showed that the sensitivity (95% CI), specificity (95%CI) and area under curve (95% CI) of the ABC score with a threshold of 0.5 or more was 0.76 (0.71 to 0.81), 0.70 (0.69 to 0.72) and 0.76 (0.73 to 0.79), respectively.

Larson

Very low quality evidence from 1 study with 5147 participants showed the sensitivity (95% CI), specificity (95% CI) and area under curve (95% CI)of the Larson score was 0.71 (0.65 to 0.76), 0.80 (0.79 to 0.81) and 0.82 (0.80 to 0.85), respectively.

McLaughlin

Very low quality evidence from one study with 372 participants showed the sensitivity (95% CI) and specificity (95% CI) of the McLaughlin score was 0.16 (0.6 to 0.31) and 0.98 (0.96 to 0 99).

Low quality evidence from one study with 396 participants showed the sensitivity, specificity and area under curve of the McLaughlin score was 0.59, 0.77 and 0.75.

Low quality evidence from one study of 596 participants showed that the area under curve of the McLaughlin score was 0.76.

Modified Field Triage Score

Very low quality evidence from 1 study of 536 participants showed that the area under curve (95% CI) of the modified Triage Score was 0.62 (0.57 to 0.67).

Prince of Wales/Rainer

Very low quality evidence from 2 studies of 2164 participants showed that the sensitivity (95% CI) Prince of Wales/Rainer score was 0.37 (0.30 to 0.44) and 0.33 (0.17 to 0.54) and the specificity (95% CI) was 0.97 (0.96 to 0.98) and 0.98 (0.97 to 0.99), respectively.

Low quality evidence from 1 study of 5147 participants showed that the sensitivity (95% CI), specificity (95% CI) and area under curve (95% CI) was 0 81 (0.76 to 0.85), 0.78 (0.77 to 0.79) and 0.86 (0.84 to 0.99), respectively.

RTS

Moderate quality evidence from 1 study of 536 participants showed that area under curve (95% CI) of the RTS score was 0.64 (0.59 to 0.69).

Schreiber

Low quality evidence from 1 study of 5147 participants showed that the sensitivity (95% CI) and specificity (95% CI) of the Schreiber score was 0.86 (0.81 to 0.90) and 0.62 (0.61 to 0.63), respectively.

TASH

Very low quality evidence from 1 study of 382 participants showed that the sensitivity (95% CI) and specificity (95% CI) of the TASH score with a threshold of 80% was 0.26 (0.13 to 0.43) and 1.0 (0.98 to 1.0), respectively.

Low quality evidence from 1 study of 596 participants showed that the area under curve of the TASH score with an unspecified threshold was 0. 84.

Very low quality evidence from 1 studies of 1030 participants showed that sensitivity (95% CI) and specificity (95% CI) of the modified TASH score with a threshold of 16 was 0.26 (0.11 to 0.46) and 0.99 (0.98 to 1.0) respectively.

Very low quality evidence from 1 study of 1134 participants showed that the sensitivity (95% CI) and specificity (95% CI) of the modified TASH score with a threshold of 18 was 0.25 (0.19 to 0.32) and 1.0 (0.99 to 1.0).

Vandromme

Low quality evidence from 1 study of 5147 participants showed that the sensitivity (95% CI), specificity (95% CI) and area under curve (95% CI) of the Vandromme score with a threshold of 1.5 or more was 0.79 (0.74 to 0.83), 0.76 (0.75 to 0.77) and 0.84 (0.82 to 0.86), respectively.

Low quality evidence from 1 study of 208 participants showed that the sensitivity and specificity of the Vandromme score with a threshold of 3 or more was 0.61 and 0.96, respectively.

Economic

No relevant economic evaluations were identified.

10.5.6. Recommendations and link to evidence

Recommendations
30.

Use physiological criteria that include the patient's haemodynamic status and their response to immediate volume resuscitation to activate the major haemorrhage protocol.

31.

Do not rely on a haemorrhagic risk tool applied at a single time point to determine the need for major haemorrhage protocol activation.

Relative values of different outcomesThe outcomes for this diagnostic review question are sensitivity and specificity of the clinical risk tools relative to a reference standard (patients receiving massive transfusion). Sensitivity is an important outcome, because poor sensitivity may result in people with potentially serious haemorrhage being undiagnosed and therefore, untreated. In contrast, low specificity, leading to incorrect positive diagnoses, will lead to unnecessary treatments (blood transfusion).
Trade-off between clinical benefits and harmsThe clinical risk tools resulted in low sensitivity and higher specificity. The GDG noted that the sensitivity and specificity of the tools were too low to be used to identify patients in need of massive transfusion. The GDG discussed the importance of clinicians having a clear set of indicators that supported them in identifying patients with life-threatening bleeding and the risks of using current risk tools .The tools need to reflect changes over time in a patient's status clearly showing responses to management rather than measuring status at a single time point. Some of the scores weight the variables which may make calculation difficult in the emergency room setting. Some of the parameters, for example laboratory analysis, may not be available within appropriate time frames.

The GDG noted the important criteria (physiological criteria, including the patient's haemodynamic status and their response to immediate volume resuscitation) in the risk tools and made a recommendation highlighting the importance of using this.
Trade-off between net health benefits and resource useNo published economic evidence was identified for this review.

Most risk scores would not necessarily have any costs associated with them directly; however, time involved in undertaking the assessment to use the score involves staff costs. Some of the scores also contain imaging or tests which would take time and involve resources.

In current practice, the tools are not commonly used and where risk tools are used, there is variation in their selection as it is felt there is no current ‘gold standard’ for predicting transfusion. This is in part due to lack of confidence in their accuracy, as well as some being difficult/time intensive to use due to the factors that need to be assessed as part of the tools. Therefore, clinical judgement tends to be used based on presenting factors, such as physiology, mechanism of injury and observation of pattern of deterioration over time (as time itself is a good diagnostic indicator).

The benefit of a risk score comes from being able to correctly stratify people (or predict outcome) to get them the right treatment. So just like a diagnostic test, there may be false negatives and positives, with people being missed or inappropriately categorised to need treatment. For the false negatives, these patients will deteriorate further, possibly leading to mortality or longer ICU stay as they are treated later when they are potentially more severe. Length of ICU stay is commonly seen as a marker of the success of strategies in predicting/identifying haemorrhage. A day in ICU for example (assuming no organs being supported) costs £619 and can increase to nearly £2000 depending on the number of organs being supported.

No clinical net benefit in using the tools was suggested by the evidence. Therefore, it is unlikely that use of any tool is cost-effective.
Quality of evidenceNine studies reported data on nine clinical risk tools.

Only studies reporting an external validation of a risk tool were included. All of the studies were retrospective cohort studies. Some of the studies had a high proportion of missing of data and many studies had wide confidence intervals. One diagnostic meta-analysis was conducted on the ABC score. This was graded as very low quality and had very serious inconsistency. 2×2 tables could not be calculated for a number of the studies and some only reported the area under the curve.
Other considerationsIn the Brockamp study, in order to compare the clinical risk tools, the AUCs were calculated and the cut-off (threshold) with the best relationship between sensitivity and specificity was used to recalculate sensitivity and specificity for each tools. This resulted in higher sensitivities than for the studies using pre-specified cut-offs. The reference standard of massive transfusion is based on physician discretion rather than biologic or laboratory outcome. Massive transfusion may have been initiated but not warranted.

The GDG did not identifying any considerations specific to children.

10.6. Intraosseous (IO)/intravenous (IV) access

10.6.1. Introduction

In trauma patients where intravenous access to provide fluids and medication is neither available nor feasible IO infusion, that is, the direct injection into the bone marrow, may be used to provide a non-collapsible entry point into the systemic venous system. It has been argued that due to the critical nature of traumatic incidences, IO access ought to be attempted in the first instance. This is because attempts at gaining peripheral access are likely to fail which can lead to a delay in treatment.

The GDG sought to identify the optimal technique for circulatory access in adults, young people and children with major trauma.

10.6.2. Review question: What is the most clinically and cost effective technique for circulatory access in patients with major trauma, including following a failed attempt at initial peripheral access?

For full details see review protocol in Appendix C.

The objectives of the clinical questions were to determine whether:

Table 64PICO characteristics of review question

PopulationChildren, young people and adults who have experienced a traumatic incident.
Intervention(s)IO
Comparison(s)IV (central and peripheral)
OutcomesCritical:
  • Mortality at 24 hours, 30 days/1 month and 12 months
  • Health-related quality of life
  • Adverse effects: pain, infection, thrombosis, multiple access failures, compartment syndrome, fracture
  • Time to establish access
Important:
  • Patient-reported outcomes (psychological wellbeing).
Study designRCTs or systematic reviews of RCTs; cohort studies that use multivariate analysis to adjust for key confounders (injury severity, age, depth of shock, degree of head injury) or were matched at baseline for these if no RCTs retrieved

10.6.3. Clinical evidence

Summary of included studies

One within-patient cohort study was included in the review 80,81; the details of which are summarised in Table 65 below. Evidence from this study is summarised in the clinical evidence summary (Table 66). See also the study selection flow chart in Appendix D study evidence tables in Appendix G, forest plots in Appendix J, GRADE tables in Appendix I and excluded studies list in Appendix K.

Table 65. Summary of studies included in the review.

Table 65

Summary of studies included in the review.

Table 66. Clinical evidence summary: IO access versus central IV access.

Table 66

Clinical evidence summary: IO access versus central IV access.

No RCTs were identified to meet the criteria for inclusion in this review. The included study was a non-randomised within-subject design study and compares IO vascular access with central venous catheterisation in patients who had failed 3 attempts at peripheral intravenous access. No comparative studies were identified for the population of patients without failed peripheral access.

10.6.4. Economic evidence

Published literature

No relevant economic evaluations were identified.

See also the economic article selection flow chart in Appendix E.

Unit costs

Below are estimates of the costs of the different methods of access. These are based on a micro costing approach of the equipment that would be needed for each type of access, based on GDG opinion.

Table 67. Resources and costs of interventions.

Table 67

Resources and costs of interventions.

Economic considerations

The study identified from the clinical review showed that the IO method takes less time to undertake, as well as leading to fewer failed attempts (a clinically important difference), thus requiring less staff time.

The difference in staff costs due to IO taking less time is shown in the table below. Only the difference in cost based on a consultant is shown here as it would be a doctor who inserts a central IV line.

Table 68. Difference in staff costs due to time taken (IO versus central IV access).

Table 68

Difference in staff costs due to time taken (IO versus central IV access).

10.6.5. Evidence statements

Clinical

Very low quality evidence from 1 non-randomised within-subject design study comprising 80 participants showed there were fewer failures to establish access in the IO access group compared with the central venous access group, with serious imprecision.

Very low quality evidence from 1 non-randomised within-subject design study comprising 80 participants showed that IO cannulation was clinically effective in terms of reducing the time to achieve vascular access, with serious imprecision.

One non-randomised within-subject design study comprising 80 participants reported no occurrence of other adverse effects, so the quality of this outcome could not be assessed.

No evidence was reported for the outcomes mortality, health-related quality of life and patient-reported outcomes (psychological wellbeing).

Economic

No relevant economic evaluations were identified.

10.6.6. Recommendations and link to evidence

RecommendationsCirculatory access in pre-hospital settings
32.

For circulatory access in patients with major trauma in pre-hospital settings:

  • use peripheral intravenous access or
  • if peripheral intravenous access fails, consider intra-osseous access.
33.

For circulatory access in children (under 16s) with major trauma, consider intra-osseous access as first-line access if peripheral access is anticipated to be difficult.

Circulatory access in hospital settings
34.

For circulatory access in patients with major trauma in hospital settings:

  • use peripheral intravenous access or
  • if peripheral intravenous access fails, consider intra-osseous access while central access is being achieved.
Relative values of different outcomesThe critical outcomes to inform decision making for this review were agreed to be mortality, health-related quality of life, time to establish access and the following adverse events; pain, infection, thrombosis, multiple access failures, compartment syndrome, fracture. These specific adverse events were chosen as the most common or potentially harmful effects of the methods of circulatory access investigated.
Time to establish circulatory access was considered a critical outcome as an accepted and well established surrogate for survival in resuscitation.

Although patient-reported outcomes, such as psychological wellbeing, including depression and anxiety, were felt to be important they were not critical to the decision making.
Trade-off between clinical benefits and harmsThe GDG agreed that the evidence in adults was quite clear in demonstrating that IO access is likely to be more rapid to achieve circulatory access for patients in whom obtaining peripheral access is not possible. It was discussed, however, that the IO route does not provide the same rate of fluid administration or drug action as large bore peripheral or central access, though no data on final outcomes was identified to support this.

Despite the lack of evidence for quality of each route of vascular access, there was consensus across the GDG that the IO route, while not sufficient for definitive circulatory access, is useful as a ‘bridging’ access technique while definitive venous access is achieved.

The benefits in using the IO route as an intermediate or ‘bridging’ route include the time to achieve access, and the minimal skill required to perform the procedure. Central venous access requires significant skill and training, however, provides definitive vascular access permitting rapid infusion of fluids and drugs.

In the absence of evidence of adverse events the GDG agreed that, in their clinical experience, adverse events from peripheral venous cannulation are not immediately life threatening. This was seen to be the case for IO access also; however, this is more painful and has additional complications associated with failed attempts (such as fracture, epiphyseal injury in children). Potential harms of central venous access were agreed to be more serious than of peripheral and IO (including haemorrhage and pneumothorax).

In view of the potential harms, it was agreed that peripheral IV access should always be considered first regardless of the age of the patient. However, there may be occasions where achieving peripheral access may be difficult (such as in infants), where IO access may be preferred as a first-line treatment. In children, IV access takes more time and can be more difficult, and the GDG felt that time is an important factor and the type of access method used should not induce delays in transfer.

It was also noted that for central access a large bore device, such as ‘swan sheath/pulmonary artery floatation catheter introducer’ is the device that should be used in patients with major trauma as they permit greater speed of fluid administration than standard central venous catheters.
Trade-off between net health benefits and resource useNo economic evidence was identified comparing IV access with IO access.

The GDG were presented with the costs of the different methods of access. For central IV access, the costs of the staff time and equipment involved in undertaking an US to guide insertion of the central IV line have not been included. It was also noted that there is no additional staff cost involved in performing the US as the same member of staff administering the line would also do this. And so all that would be involved would be the opportunity cost of the doctor's time and the equipment. It was also highlighted that IO access is only used as a bridging method, with the ultimate goal being central IV access.

Adverse events were also discussed and should be considered within the trade-off. These include pain being a big factor for IO access in conscious patients, not just for insertion of the needle but also while fluids are being inserted. Additional adverse events can include fracture. For IV access adverse events can include infection and misplacement.

The clinical review data showed that IO access takes less time to perform, and also leads to fewer failed attempts (clinically significant difference), although the study was unclear as to what categorised as a ‘failed attempt’. The data was also of a low quality.

The GDG felt that the difference in time taken between the two methods was a clinically-relevant factor in a time critical situation when you are trying to get fluids into a patient as quickly as possible for resuscitation purposes.

As mentioned above, the US staff time costs and equipment have not been included; therefore, it is unclear if central IV access is likely to be more expensive than IO access. However, given that IO is used as a bridging method for inserting fluids until central IV access can be performed, then IO access is the more expensive option as this will include the IO access followed by central IV access. Using IO access while central access is being achieved was considered cost effective as it would ensure quicker access. The GDG felt that time is an important factor and the type of access method used should not induce delays in transfer, as time is a critical factor in major trauma patients and delay can impact patient outcomes.
Quality of evidenceOne single study was identified comparing IV with IO access. The study population was seriously ill or critically injured adult patients (requiring fluid resuscitation) who had received three failed attempts at peripheral large-bore cannula insertion. It was unclear precisely whether this population was entirely a trauma population, therefore, the evidence was somewhat indirect.

The design of this study was non-randomised within-patient, which, while eliminating the bias arising from baseline differences, also made any meaningful comparison of the effect of the equipment impossible. The only outcomes reported were time to achieve access and unsuccessful attempts.

Given the small size of the study sample, there was large variation and so considerable uncertainty around the effect estimate, for which the quality of the evidence was downgraded accordingly.
Other considerationsThe GDG noted that if achieving peripheral access is difficult, further attempts should not delay transport to hospital. IV access is often not absolutely required. If circulatory access is required, IO access is rapid and where appropriate should be recommended in the pre-hospital environment.

Circulatory access in children (particularly infants and patients who are cold) may be more difficult. In the pre-hospital environment, IO access, where appropriate, may be the first-line preferred method for circulatory access.

10.7. Volume resuscitation

10.7.1. Introduction

Uncontrolled bleeding remains the leading cause of preventable death following major trauma and requires early detection and prompt action to resuscitate the patient with fluid and achieve a stable hemodynamic status. Traditionally, fluid resuscitation of an actively bleeding patient has emphasised the maintenance of normal circulation in order to maintain organ perfusion. However, studies have indicated that limiting the amount of fluids administered using permissive hypotension during the initial resuscitation period may improve trauma outcomes. However, the evidence for the practice remains limited and practice may differ depending on type of injury (penetrating or blunt). Moreover, much of the evidence was made before the use of haemostatic resuscitation and clear guidance on resuscitation strategy is still required.

10.7.2. Review question: What are the most clinically and cost effective fluid resuscitation strategies in the major trauma patient (hypotensive versus normotensive)?

For full details see review protocol in Appendix C.

Table 69PICO characteristics of review question

PopulationChildren, young people and adults experiencing a traumatic incident with acute haemorrhage.
Intervention:Combination of permissive hypotension and normotension
Permissive hypotension
ComparisonResuscitation with normotension as aim
OutcomesCritical:
  • Mortality at 24 hours, 30 days/1 month, and 12 months
  • Health-related quality of life
  • Neurological outcome
  • Length of intensive care stay
  • Blood product use
Important:
  • Multi-organ failure
  • Time to definitive control of haemorrhage
  • Patient-reported outcomes: pain/discomfort return to normal activities psychological wellbeing)
Study designRCTs or systematic reviews of RCTs; cohort studies that use multivariate analysis to adjust for key confounders (injury severity, age, depth of shock, degree of head injury) or were matched at baseline for these if no RCTs retrieved

10.7.3. Clinical evidence

Two studies were included in the review; 13,41 these are summarised in Table 70 below. Evidence from these studies is summarised in the clinical evidence summary. See also the study selection flow chart in Appendix D study evidence tables in Appendix G, forest plots in Appendix J, GRADE tables in Appendix I and excluded studies list in Appendix K.

Table 70. Summary of studies included in the review.

Table 70

Summary of studies included in the review.

Bickell13 considered fluid resuscitation strategies in the pre-hospital and Dutton41/within hospital population. Studies were also stratified by mechanism of injury with two studies 13 analysed for penetrating trauma. No studies were found in populations with an exclusively blunt mechanism of injury and in children and young people.

Table 71. Clinical evidence summary: Permissive Hypotension versus Resuscitation with normotension as aim – Pre-hospital.

Table 71

Clinical evidence summary: Permissive Hypotension versus Resuscitation with normotension as aim – Pre-hospital.

Table 72. Clinical evidence summary: Permissive Hypotension versus Resuscitation with normotension as aim – In-hospital.

Table 72

Clinical evidence summary: Permissive Hypotension versus Resuscitation with normotension as aim – In-hospital.

Table 73. Clinical evidence summary: Permissive Hypotension versus Resuscitation with normotension as aim – In-hospital (Combined).

Table 73

Clinical evidence summary: Permissive Hypotension versus Resuscitation with normotension as aim – In-hospital (Combined).

Table 74. Clinical evidence summary: Permissive Hypotension versus Resuscitation with normotension as aim – Penetrating Injury.

Table 74

Clinical evidence summary: Permissive Hypotension versus Resuscitation with normotension as aim – Penetrating Injury.

10.7.4. Economic evidence

Published literature

One economic evaluation relating to this review question was identified but excluded due to a combination of limited applicability and methodological limitations.136,136 This is listed in Appendix L, with reasons for exclusion given.

See also the economic article selection flow chart in Appendix E.

Unit costs

Below are the costs of the various fluid and blood products available, to aid consideration of cost effectiveness.

Table 75. Fluid and blood product costs.

Table 75

Fluid and blood product costs.

As well as the costs of the products themselves, additional costs would be involved in the handling and administration of the products, which would apply each time a product is issued, and could vary depending on the type of product (see more in section 10.8.4).

10.7.5. Evidence statements

Clinical

Permissive Hypotension versus normotension – Pre-hospital

Moderate quality evidence from a 1 RCT comprising of 598 participants demonstrated a clinical benefit of permissive hypotension over normotension for mortality at 30 days, with serious imprecision.

High quality evidence from 1 RCT with 598 participants demonstrated no clinical difference between permissive hypotension and normotension for length of ICU stay, with no serious imprecision.

Low quality evidence from 1 RCT comprising of 598 participants demonstrated a clinical benefit of permissive hypotension over normotension for multi-organ failure at 30 days, with serious imprecision.

Permissive Hypotension versus normotension – In hospital

Low quality evidence from 1 RCT comprising of 110 participants demonstrated no clinical difference between permissive hypotension and normotension for mortality at 24 hours, with very serious imprecision.

Low quality evidence from 1 RCT comprising of 110 participants demonstrated no clinical difference between permissive hypotension and normotension for mortality at 30 days, with very serious imprecision.

Moderate quality evidence from 1 RCT comprising of 110 participants demonstrated no clinical difference between permissive hypotension and normotension for time to definitive control of haemorrhage, with serious imprecision.

Permissive Hypotension versus normotension – In hospital (Combined)

Very low quality evidence from 2 RCTs with 708 participants demonstrated a clinical benefit of permissive hypotension over normotension for mortality at 24 hours, with serious imprecision.

Very low quality evidence from 2 RCTs with 708 participants demonstrated a clinical benefit of permissive hypotension over normotension for mortality at 30 days, with serious imprecision.

High quality evidence from 1 RCT with 598 participants demonstrated no clinical difference between permissive hypotension and normotension for ICU length of stay, with no imprecision.

Low quality evidence from 1 RCT comprising of 598 participants demonstrated a clinical benefit of permissive hypotension over normotension for multi-organ failure at 30 days, with serious imprecision.

Moderate quality evidence from 1 RCT comprising of 110 participants demonstrated no clinical difference between permissive hypotension and normotension for time to definitive control of haemorrhage, with serious imprecision.

Permissive Hypotension versus Normotension – Penetrating Injury

Moderate quality evidence from 1 RCT comprising of 598 participants demonstrated a clinical benefit of permissive hypotension over normotension for Mortality at 24 hours, with serious imprecision.

High quality evidence from 1 RCT with 598 participants demonstrated no clinical difference between permissive hypotension and normotension for length of ICU stay, with no imprecision.

Low quality evidence from 1 RCT comprising of 598 participants demonstrated a clinical benefit of permissive hypotension over normotension for multi-organ failure at 30 days, with serious imprecision.

Economic

No economic evidence identified

10.7.6. Recommendations and link to evidence

Recommendations
35.

For patients with active bleeding use a restrictive approach to volume resuscitation until definitive early control of bleeding has been achieved.

36.

In pre-hospital settings, titrate volume resuscitation to maintain a palpable central pulse (carotid or femoral).

37.

In hospital settings, move rapidly to haemorrhage control, titrating volume resuscitation to maintain central circulation until control is achieved.

38.

For patients who have haemorrhagic shock and a traumatic brain injury:

  • if haemorrhagic shock is the dominant condition, continue restrictive volume resuscitation or
  • if traumatic brain injury is the dominant condition, use a less restrictive volume resuscitation approach to maintain cerebral perfusion.
Relative values of different outcomesMortality, health-related quality of life, length of intensive care stay, neurological complications and blood product use were critical outcomes. Multi-organ failure, time to definitive control of haemorrhage, patients reported outcomes, return to normal activities and psychological wellbeing were important outcomes. The selection of outcomes reflects both short-term and long-term sequelae of fluid replacement therapy.
Trade-off between clinical benefits and harmsPre-hospital
Pre-hospital evidence from a single study on adults indicated that a permissive hypotension strategy was associated with reduced mortality at 30 days and multi-organ failure. The GDG discussed the concept of permissive hypotension and noted that the majority of evidence supporting this practice has been conducted in animal studies and demonstrated that liberal fluid resuscitation reduces in vivo coagulation.

Based on the evidence, their clinical experience and the fact that a restrictive approach to fluid administration in patients with trauma (particularly those who are actively bleeding) is the current practice, the GDG recommended a restrictive approach to fluid administration (that is, small boluses of crystalloid or blood components).

The GDG then discussed how to titrate resuscitation to a specific target (that is, 50 mmHg arterial pressure for permissive hypotension). The GDG noted that identification of active bleeding and shock is difficult. Moreover, recordings of blood pressure may be inaccurate in the critically ill patient and measurement difficulties are intensified in the pre-hospital setting.

The GDG discussed the various indicators of shock, but felt that a simple assessment tool, such as assessment of central pulse (carotid or femoral), would be more reliable for the pre-hospital clinician and allow patients to be transported quicker for definitive care. The GDG also discussed the measurement of radial pulse, but felt a central indicator of pulse matched the blood pressure targets used in the clinical studies (pre and hospital). The central pulse is also easier to palpate than a radial pulse.

The GDG also noted that the patients response following initial fluid bolus, would provide the best information if this patient is shocked and ongoing management should be based on this (that is, a patient who does not respond to initial volume resuscitation is likely to be actively bleeding).

In hospital
The GDG discussed the evidence for in-hospital resuscitation from a small single study in adults. The study demonstrated no difference between normotension and permissive hypotension for mortality and time to definitive control of haemorrhage. However, it was felt that the study by Bickell et al. could be combined with this study (as the intervention is also carried out in hospital). When combined, the evidence suggested a benefit of a permissive strategy (systolic blood pressure titrated to 70 mmHg and delayed resuscitation) for mortality at 24 hours and 28 days.

The GDG noted the benefits of permissive hypotension in the emergency department and physiological rationale, which was similar to the pre-hospital. However, the GDG felt this should be governed by a rapid protocol to definitively stop the source haemorrhage, as this is a critical factor for improved patient outcome.

The GDG discussed titration of the resuscitation strategy to a set BP, but the evidence from both studies demonstrated this to be difficult. The GDG, therefore, agreed that this should be titrated against maintenance of central circulation, as there would be risk of organ failure and death if central circulation was unable to maintain vital organ perfusion. The measurement of blood pressure may be more reliable (for example, with the placement of an arterial line) in hospital.

Head injury
The GDG discussed a subpopulation of patients with traumatic brain injury in which hypotension may be associated with worse outcome. No clinical evidence was presented for patients with combined head injury and active bleeding and the GDG noted that majority of practice in this area was based on animal studies. The GDG, therefore, made a recommendation on these patients using informal consensus. Assessment of severity of head injury is difficult in the pre-hospital setting. The GDG noted that patients with a severe brain injury may not benefit from a restrictive fluid approach and that maintenance of cerebral perfusion best reflects current practice.

The GDG also noted that in patients who present with both traumatic brain injury and active bleeding, the physician should treat with an appropriate strategy for the most severe injury (that is, use a restrictive approach if active bleeding is the prevailing injury and brain injury secondary).
Trade-off between net health benefits and resource useNo economic evidence was identified for this question.

Permissive hypotension involves delaying or restricting fluids in order to reach a defined endpoint – which is lower than normal blood pressure. Normotension as an aim involves giving fluids until the patients has reached the normal level of blood pressure (pre-injury levels).

The types of fluids used would have an impact on costs, as blood is comparably more expensive than crystalloids. It would be assumed that the normotensive group would receive more fluids overall than the permissive group.

There are various adverse events that could be associated with the different techniques and it is often unclear which method is the best. It is perceived that normotensive hypotension can lead to more bleeding and have a negative effect on the clotting process by raising blood pressure to a higher level.

The clinical evidence tended to favour permissive hypotension, particularly the larger study (Bickell et al.), which reported lower mortality, length of stay, and fewer cases of multiple organ failure for the permissive group. Thus potentially a dominating strategy.

This recommendation will lead to a change in practice as currently, clinicians are trained to follow an approach of early recognition and intervention. The recommendation may be cost saving as fewer units of fluids will be used.
Quality of evidenceThe GDG noted that clinical data from well controlled, prospective trials applying the concept of permissive hypotension in trauma patients (as described in the protocol) was still missing.

Pre-hospital
The GDG noted that the single study comparing hypotension with normotension was a well conducted single centre RCT. However, the GDG noted that crystalloids were used, which are not recommended over haemostatic resuscitation agents in our guidance. The outcome of multi-organ failure was not reported clearly and said to be at an increased risk of bias. Overall, the outcomes were rated from high to low quality.

The GDG noted that the objective of that study was the comparison between standard pre-hospital trauma fluid resuscitation versus delayed onset of fluid resuscitation (fluid not administered until patients reached the operating room). This did not directly compare a permissive hypotension strategy with normotension, but the GDG noted that the restrictive approach to fluid replacement (as described in the study) is commonly used in UK pre-hospital practice and detailed in consensus guidelines, such as the Joint Royal Colleges Ambulance Liaison Committee. Moreover, the GDG noted that only patients with penetrating trauma were included in the study and these patients would require a more conservative fluid management.

In hospital
The GDG noted that the small RCT was conducted in a mixed population with both blunt and penetrating injuries. The blood pressure following the intervention was similar and the subjects were not truly randomised to each respective arm. However, the GDG noted that this may due to physiological adaption. All outcomes were rated at a high risk of bias.

The GDG also discussed that the Bickell et al. study could be interpreted to cover in-hospital and pre-hospital, but the overlap is not clear in the paper and presents an additional risk of bias. All outcomes combining both studies were rated a very high risk bias.
Other considerationsThe GDG noted that there were varying interpretations of the meaning and goals of permissive hypotension and that the technique used to achieve permissive hypotension varied across the studies and in practice.

The GDG also noted that local practice will be based on proximity to a trauma centre; whereby it is essential resuscitate a patient in shock if there is a prolonged travel time to definitive management.

The GDG did not identify any considerations specific to children.

10.8. Fluid replacement

10.8.1. Introduction

The main purpose of volume replacement in traumatized patients is to re-establish tissue perfusion. There is not yet a consensus about which if any fluids should be used in trauma patients.

10.8.2. Review question: What is the best volume expansion fluid to use in the resuscitation of haemorrhagic shock?

For full details see review protocol in Appendix C.

Table 76PICO characteristics of review question

PopulationChildren, young people and adults who have experienced a traumatic incident.
Intervention(s)
  • Red blood cells (RBCs)
  • Fresh frozen plasma (FFP)
  • Liquid plasma
  • Crystalloids
  • Lyophilised plasma
Comparison(s)A comparison or combination of the above (including different ratios)
OutcomesCritical:
  • Mortality at 24 hours, 30 days/1 month and 12 months
  • Health-related quality of life
  • Length of intensive care stay
  • Acute transfusion reaction:
    • haemolytic transfusion reaction – acute
    • haemolytic transfusion reaction – delayed
    • post-transfusion purpura
    • previously uncategorised complications of transfusion
    • transfusion-associated graft versus host disease
    • transfusion-associated circulatory overload
    • transfusion associated dyspnoea
    • transfusion-related acute lung injury
    • transfusion-transmitted infections
Important:
  • Time to definitive control of haemorrhage
  • Patient-reported outcomes:
    • return to normal activities
    • psychological wellbeing
Study designRCTs or systematic reviews of RCTs; cohorts if no RCTs retrieved (matched at baseline or adjusted for age, Glasgow coma scale [GCS], injury severity score and shock)

10.8.3. Clinical evidence

One RCT comparing plasma, platelet and RBCs in a ratio of 1:1:1 to 1:1:263,64 and one secondary analysis of a prospective cohort study was identified comparing different ratios of crystalloid with RBCs 102,103 and one RCT comparing two different crystalloids143,144 were identified. The study comparing different ratios of plasma, platelet and RBCs was included even though one of the products was platelets because the GDG were confident that the addition of platelets to plasma and RBCs was not likely to influence the ratio of plasma to RBCs used. Also, platelets are frequently given with volume expansion products to control coagulopathy. Both studies were in the civilian population. The study characteristics are summarised in Table 77 below. Evidence from these studies is summarised in the clinical evidence summary below (Table 78). See also the study selection flow chart in Appendix D study evidence tables in Appendix G, forest plots in Appendix J, GRADE tables in Appendix I and excluded studies list in Appendix K.

Table 77. Summary of studies included in the review – Massive transfusion by risk of survivor bias.

Table 77

Summary of studies included in the review – Massive transfusion by risk of survivor bias.

Table 78. Clinical evidence summary: FFP: platelets: RBCs.

Table 78

Clinical evidence summary: FFP: platelets: RBCs.

Only studies on patients undergoing massive transfusion were included.

Table 79. Clinical evidence summary: Crystalloid: RBCs.

Table 79

Clinical evidence summary: Crystalloid: RBCs.

Table 80. Clinical evidence summary: Crystalloid (0.9%NaCl) versus crystalloid (Plasma-Lyte A).

Table 80

Clinical evidence summary: Crystalloid (0.9%NaCl) versus crystalloid (Plasma-Lyte A).

Narrative review

The study comparing plasma:platelets:RBCs 1:1:1 versus 1:1:263,64 reported the median (IQR) ICU-free days as 5 (0 to 11) versus 4 (0 to 10) (0=0.10). The median (IQR) GCS – Extended for 1:1 versus 1:1:2 was 4 (3 to 6) (n=30) versus 4.5 (3.5 to 7.0) (n=30) (p=0.11).

When a dose-response relationship was evaluated in the study comparing ratios of crystalloid:pRBC 102,103, regression analysis revealed that a crystalloid:pRBC ratio was associated with: multiple organ failure (OR 2.6; 95%CI 1.2 to 5.4, p=0.011) and acute respiratory distress syndrome (OR 2.5; 95%CI 1.2 to 4.9; p=0.010).

The study comparing 0.9% NaCl with Plasma-Lyte A 143,144reported a median (IQR) hospital length of stay of 9 (4, 30) versus 12 (4, 21) days and an ICU length of stay of 4 (2, 13) versus 4 (1, 9) days.

10.8.4. Economic evidence

Published literature

No economic evidence identified.

See also the economic article selection flow chart in Appendix E.

Unit costs

Table 81. Blood product costs.

Table 81

Blood product costs.

Note that for children, FFP is substantially more expensive due to the Department of Health recommendations that those born after 01.01.96 should use particular types of FFP and cryoprecipitate that have undergone additional reduction procedures to reduce the risk of viruses. Please see chapter 10.4.4 for more detail on this.

Lyophilised plasma and liquid plasma are not as commonly used in practice as fresh frozen plasma. Costs could not be sourced for these interventions.

Issues of finite supply and wastage arise in the use of blood components. In order to avoid waiting for cross matching tests, patients are mostly given the universal donor blood type, which may be in short supply. This is relevant for both blood and plasma. Additionally with FFP, this takes time to thaw. Some major trauma centres may have a small amount of plasma pre-thawed which can save time in a time critical situation, however, pre-thawed plasma only has a shelf life of 24 hours, and thus, is wasted if a suitable patient is not identified to use it on.

As well as the costs of the products themselves, there are also administration costs involved from the hospital laboratory that must prepare and issue the products. For example, the following costs are from GDG member contact and illustrate the blood product handling and administration costs involved in providing the blood components:

Table 82. Blood product handling and administration costs.

Table 82

Blood product handling and administration costs.

10.8.5. Evidence statements

Clinical

Plasma: platelets: RBCs

Moderate quality evidence from 1 RCT comprising 680 participants showed that a 1:1:1 ratio was clinically effective compared with 1:1:2 in terms of mortality (24 hours, 30 days), with serious imprecision.

Low and Very low quality evidence from 1 RCT comprising 680 participants showed there was no difference in clinical harm between a1:1:1 and a 1:1:2 ratio in terms transfusion-related metabolic complication and transfusion-associated circulator overload, with very serious imprecision.

High quality evidence from 1 RCT comprising 680 participants showed that a 1:1:1 ratio was clinically effective compared with 1:1:2 in terms of achieving haemostasis, with no imprecision.

Moderate quality evidence from 1 RCT comprising 680 participants showed that a 1:1:1 ratio was clinically effective compared with 1:1:2 in terms of discharged home, with serious imprecision.

Crystalloids: RBCs

Very low quality evidence from 1 cohort study comprising 452 participants showed there was no difference in clinical effectiveness between a high and a low ratio in terms mortality, with very serious imprecision.

Very low quality evidence from 1 cohort comprising 452 participants showed there was no difference in clinical harm between a high and a low ratio in terms nosocomial infection, with very serious imprecision.

Very low quality evidence from 1 cohort comprising 452 participants showed that a high ratio was clinically harmful compared with low in terms of multiple organ failure, with serious imprecision.

Very low quality evidence from 1 cohort comprising 452 participants showed that a high ratio was clinically harmful compared with low in terms of multiple acute respiratory distress syndrome, with no imprecision.

Economic

No economic evidence identified

10.8.6. Recommendations and link to evidence

Recommendations
39.

In pre-hospital settings only use crystalloids to replace fluid volume in patients with active bleeding if blood components are not available.

40.

In hospital settings do not use crystalloids for patients with active bleeding (See the section on resuscitation in the NICE guideline ‘Intravenous fluid therapy in adults in hospital’ and the section on fluid resuscitation in the NICE guideline ‘Intravenous fluid therapy in children and young people in hospital’ for advice on tetrastarches.

41.

For adults (16 or over) use a ratio of 1 unit of plasma to 1 unit of red blood cells to replace fluid volume.

42.

For children (under 16s) use a ratio of 1 part plasma to 1 part red blood cells, and base the volume on the child's weight.

Relative values of different outcomesMortality, health-related quality of life, length of intensive care stay and transfusion-related complications were critical outcomes. Time to definitive control of haemorrhage, patient-reported outcomes, return to normal activities and psychological wellbeing were important outcomes. The selection of outcomes reflects both short-term and long-term sequelae of fluid replacement therapy.

For the RCT comparing plasma:platelets:RBCs 1:1:1 versus 1:1:2 the only outcomes reported were in-hospital mortality, transfusion-related metabolic complications, transfusion-associated circulatory overload, achieved haemostasis and discharged home. For the study comparing high versus low ratio of crystalloids with RBCs, the only outcomes reported were in-hospital mortality, nosocomial infection, multiple organ failure and acute respiratory distress syndrome. For the RCT comparing two different types of crystalloids, the only outcome reported was mortality.
Trade-off between clinical benefits and harmsThe RCT in adults comparing on FFP:platelets:RBCs reported of clinically important reduction in mortality at 24 hours and 30 days in favour of a 1:1:1 ratio compared with a 1:1:2 ratio but with serious imprecision. Clinically important benefits were also reported with 1:1:1 for the number of patients achieving haemostasis (with no imprecision) and the number of patients discharged home (with serious imprecision). No clinically important harms were associated with 1:1:1 (transfusion-related related metabolic complications and transfusion-related circulatory overload).

One cohort study on adults evaluating different ratios of crystalloids with RBCs, reported no difference in in-hospital mortality, but there was evidence that higher ratios were associated with significant harms (nosocomial infection, multiple organ failure and acute respiratory distress syndrome).

One RCT on adults compared two different types of crystalloid administered in-hospital. It reported a higher mortality rate with 0.9% NaCl compared with Plasma-Lyte A (but the difference was not considered to be clinically important), but with very serious imprecision.

The GDG acknowledged that a recommendation to avoid using crystalloids and other clear fluids except in patients with profound haemorrhagic shock in the pre-hospital environment is a change in clinical practice. The GDG wanted to highlight that haemorrhage and other forms of shock (inadequate perfusion of end organs) in major trauma has a potential early and continued detrimental effect on clotting function ranging from an alteration in the complex systems involved in clotting itself to an absolute reduction in the body's raw materials required for creating adequate clot formation.

Both crystalloids and colloids have an effect upon the complex clotting systems and their effective function in the patient who is severely injured. Additionally, continued and prolonged periods of shock have a detrimental effect upon outcome manifesting as inadequate perfusion of organs converting to organ failure and hence multi-organ dysfunction syndrome – so there is impact from the severity of the shock, the type of shock, the length of time that the patient is shocked for; this will affect the end organs and the clotting systems and both (for example, bone marrow and haematopoietic organs and their capability to manufacture essential ingredients for clot formation and replenishment of the circulating blood components and volume). The optimum management is fluid replacement with blood components.

The GDG discussed the situation when a pre-hospital practitioner is treating a patient in profound haemorrhagic shock but does not have access to blood components. In this case small boluses of crystalloids would be appropriate.
Trade-off between net health benefits and resource useNo economic evidence was identified for this question.

The different blood components can vary in price with blood being the most expensive of those currently used in practice (lyophilised plasma and liquid plasma are not commonly used in practice). Costs will also increase substantially if many units are used. The high ratio strategies will have a higher cost as there is more of the first product being used for every unit of the second (for example, 1:1 [high ratio] versus 1:2 [low ratio]). The cost effectiveness of the higher units is thus dependent on their effectiveness in resuscitating patients. Blood product wastage and the opportunity cost of not being able to use these blood components on others are also costs that need to be taken into account. Blood components also have administration costs associated with the handling and issuing of the products from the hospital laboratory.

If blood was to be used pre-hospital then this would lead to additional costs of storage equipment and clinical systems and waste of unused blood if all pre-hospital services had to start carrying blood or blood components.

There are various complications associated with transfusion which can lead to long-term morbidities or mortality. Although the risk of transfusion-related complications is likely to be low, infection and the immunological consequences are more common. For more information on these risks please see the latest SHOT report (ref), an extract of which can be found in Appendix O.

The evidence showed that high ratios of FFP to blood are favoured (more FFP per unit of blood), and low ratios of crystalloids to blood are favoured (less crystalloids per unit of blood). Although, this evidence was of low quality.

The GDG felt that recommending a balanced administration of FFP and RBCs, with a ratio of 1:1, was appropriate given the evidence. Current practice is variable with regards to ratios used, however, 1:1 is likely to be more common. The supply of FFP is a consideration here because a ratio of 1:1 in comparison to 1:2 means twice as much FFB will be needed per unit of blood. This creates additional pressure on the limited supply of the universal donor group AB, and also adds further pressure to have blood tests done earlier in order to match the plasma to the patients' blood type if group AB is not available. Wastage is also an issue because if more plasma needs to be used then having more plasma de-thawed which has only a 24-hour shelf life may lead to more wastage problems and high opportunity cost. FFP for children is over 6 times more expensive than that of adults because department of Health recommendations state that those born after 1996 must use specially treated FFP. Therefore, this recommendation is likely to have an effect on practice and a cost impact.

A statement about crystalloids and other clear fluids was made, with awareness that blood may not be available pre-hospital. Avoiding the use of crystalloids pre-hospital is a change to current practice which could potentially result in cost savings if crystalloids are not used, however, these are relatively cheap, and this saving may be offset if this recommendation encourages blood use pre-hospital. In hospital, if clear fluids are avoided, this also increases the pressure to have blood available as soon as possible. No comment could be made on the other interventions in the protocol as no evidence was identified and also these are not commonly used in practice.
Quality of evidenceThe RCT had no methodological limitations but evidence was downgraded for imprecision where appropriate. The quality of the evidence ranged from low to high quality.

Only one study was identified comparing different ratios of crystalloids with RBCs and this had been adjusted for survivor bias by excluding patients who died within 24 hours' of admission. The outcomes were graded as very low quality due to study design and imprecision.

The RCT comparing two different types of crystalloids reported one low quality outcome due to imprecision.
Other considerationsThe GDG noted that the patients in the 1:1:1 arm received platelets first (6 units) followed by alternating RBCs and plasma. In comparison, in the 1:1:2 arms patients received two units of RBCs first and one unit of plasma. Platelets were not transfused until after nine units of other blood components. The total number of platelets received in the 1:1: arm was slightly higher than it should have been (if given in accordance with a 1:1:1 ratio) and slightly lower in the 1:1:2 arm.

Despite the considerations above, the GDG felt that a 1:1 ratio of plasma to RBCs is beneficial especially in the absence of any clinical harm. This ratio is the most similar in composition to whole blood.

The recommendation to administer small boluses of crystalloids and to avoid other clear fluids is specific to the major trauma patients who are actively bleeding. To note, the NICE guideline on intravenous fluid therapy in adults is specific to hospitalised patients and excluded the major trauma population.

With the exception of adjusting dose to weight, the GDG identified no consideration specific to children.

The GDG noted the there are people who refuse blood components based on religious beliefs. This was not a problem specific to Major Trauma and the issue occurs in other clinical situations. The GDG noted NG24 Blood transfusion includes recommendations that address this issue. The committee acknowledged there may be unique situations in Major trauma when a patient is unable to give consent (either they are unconscious or have reduced mental capacity) and these situations are dealt with on an individual basis.
Copyright © National Clinical Guideline Centre, 2016.
Bookshelf ID: NBK368100

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (3.7M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...