Medical anti‐shock trousers (pneumatic anti‐shock garments) for circulatory support in patients with trauma

Abstract

Background

Medical antishock trousers (MAST) have been used to increase venous return to the heart until definitive care could be given. This, combined with compression of blood vessels, is believed to cause the movement of blood from the lower body to the brain, heart and lungs. However, the equipment is expensive, and may have adverse effects.

Objectives

To quantify the effect on mortality and morbidity of the use of medical anti‐shock trousers (MAST)/pneumatic anti‐shock garments (PASG) in patients following trauma.

Search methods

Trials were identified by searches of the Cochrane Injuries Group Specialised Register (April 2007), the Cochrane Central Register of Controlled Trials (Issue 1, 2007), MEDLINE (April 2007), EMBASE (April 2007), ISI Web of Science (April 2007), National Research Register (Issue 1, 2007) and PubMed (April 2006). In addition we checked the reference lists of relevant trials and reviews. We contacted current researchers in the field for unpublished data and ongoing trials.

Selection criteria

Randomised and quasi‐randomised trials of MAST/PASG in patients following trauma (excluding fractures of the extremities in which MAST/PASG may be used as a splint).

Data collection and analysis

Data were extracted independently by two authors. Data were collected on mortality, duration of hospitalisation and ICU stay, and quality of allocation concealment.

Main results

Two trials were identified that met the inclusion criteria. These trials included 1202 randomised patients in total; however, data for 1075 of these were available. The relative risk of death with MAST was 1.13 (95% CI 0.97 to 1.32). Duration of hospitalisation and of intensive care unit stay was longer in the MAST treated group. The weighted mean difference in the length of intensive care unit stay was 1.7 days (95% CI 0.33 to 2.98).

Authors' conclusions

There is no evidence to suggest that MAST/PASG application reduces mortality, length of hospitalisation or length of ICU stay in trauma patients and it is possible that it may increase these. These data do not support the continued use of MAST/PASG in the situation described. However, it should be recognised that, due to the poor quality of the trials, conclusions should be drawn with caution.

Plain language summary

No evidence that medical anti‐shock trousers improve survival for people with trauma, and some suggestion they may worsen outcomes, but more trials are needed

About one third of injury deaths are due to shock from blood loss. Preventing shock in people with uncontrolled bleeding is therefore vital. Treatment aims to maintain blood pressure, so that tissue damage is minimised. Medical anti‐shock trousers (MAST) are believed to increase blood pressure and blood flow to the heart and brain, helping to stabilise the person until they receive further treatment. The review of trials found no evidence that MAST application decreases deaths, with some suggestion that it may even do harm. More research is needed.

Background

In 1903, Crile described his "pneumatic rubber suit", a device designed to decrease postural hypotension in neurosurgical patients (Crile 1903). Crile's design was later used in World War II to prevent blackout of pilots subjected to high G forces. During the Vietnam War the suit was used in order to stabilise patients with haemorrhagic shock during transportation (Schwab 1983). In 1973, MAST (Pneumatic Anti‐shock Garments) were developed for clinical use in the treatment of haemorrhagic shock or hypotension in emergency prehospital and emergency department settings (Davis 1986). MAST were used to increase venous return to the heart until definitive care (transfusion and/or surgical intervention) could be given. MAST was thought to cause autotransfusion of blood from the lower extremities and an increase in systemic vascular resistance. This, combined with compression of blood vessels, is believed to cause the movement of blood from the lower body to the brain, heart and lungs (Jennings 1986). In 1977, the Committee on Trauma of the American College of Surgeons included anti‐shock trousers on the list of essential equipment for ambulances (Schwab 1983). However, the equipment is expensive (Schneider 1989), and may have adverse effects.

We conducted a systematic review of randomised controlled trials to quantify the effect on mortality and morbidity of the use of MAST in patients following trauma.

Objectives

To quantify the effectiveness and safety of Medical Anti‐Shock Trousers (Pneumatic Anti‐Shock Garments) in patients following trauma.

Methods

Criteria for considering studies for this review

Types of studies

Randomised and quasi‐randomised controlled trials comparing the use of MAST/PASG to a control group not given this treatment.

Types of participants

Patients following trauma (fractures of the extremities in which MAST/PASG may be used as a splint are excluded).

Types of interventions

Application of MAST/PASG.

Types of outcome measures

Outcome measures of interest were:

• Mortality

• Length of hospital stay

• Length of time spent in Intensive Care Unit (ICU).

Search methods for identification of studies

Electronic searches

Initial searches of electronic databases were carried out in January 1999. Original search strategies can be view in the additional tables (Table 1; Table 2).

1. MEDLINE (1966 to 1999; SilverPlatter ERL Version).

1. (anti shock or antishock) trouser* 2. G‐suits [MeSH] 3. MAST and (shock or antishock or trouser*) 4. PASG and (shock or antishock or trouser*) 5. 1 or 2 or 3 or 4

2. EMBASE (1980 to 1999; SilverPlatter ERL Version).

1. (anti shock or antishock) trouser* 2. MAST‐suit [EMBASE keyword] 3. MAST and (shock or antishock or trouser*) 4. PASG and (shock or antishock or trouser*) 5. 1 or 2 or 3 or 4

The most recent searches were conducted in April 2007, and the following databases were searched:

• MEDLINE (Ovid) (to April 3 2007)

• EMBASE (Ovid) (to April 3 2007)

• Central (to 2007, Issue 1)

• NRR (to 2007, Issue 1)

• PubMed (searched on April 3 2007 (last 6 months))

• ISI Web of Science (searched 4 April 2007)

The search strategies are in Appendix 1.

Trials 'filters' based on those developed by Glanville 2006 were used to restrict search results.

Searching other resources

We checked the reference lists of relevant trials and reviews. We contacted current researchers in the field for unpublished data and ongoing trials. The searches were not restricted by language or publication status.

Data collection and analysis

Selection of studies

One author (KD) examined the electronic search results to identify reports of possibly relevant trials, which were then retrieved in full. Two authors (KD, IR) applied selection criteria to the trial reports and resolved disagreements by discussion.

Data extraction and management

One author (KD) extracted information on the following:

• method of allocation concealment,

• number of randomised patients,

• types of participants and the interventions,

• outcome data (number of deaths, the length of hospitalisation, and the duration of any ICU stay).

The author was not blinded to the authors or the journal when doing this.

Assessment of risk of bias in included studies

Allocation concealment quality has been shown to effect the results of studies (Schulz 1995). Two authors (KD, IR) scored this quality on the scale used by Schulz 1995, as shown below (C is assigned to the poorest quality and A to the best quality allocation concealment).

• C = trials with inadequate concealment (for example, alternation, reference to case record numbers or dates of birth).

• B = trials in which allocation concealment approach was not reported or in which an approach was reported that did not fall into the other categories.

• A = trials deemed to have taken adequate measures to conceal allocation (for example, central randomisation; numbered or coded bottles or containers; drugs prepared by the pharmacy; serially numbered, opaque, sealed envelopes; or other description that contained elements convincing of concealment).

In order to perform an intention‐to‐treat analysis, information regarding loss to follow‐up was sought by contacting the trial authors. Information on blinding (of the outcome assessor) was not sought as it was not relevant for these outcomes.

Data synthesis

The data from the two trials were considered sufficiently homogeneous for the data to be pooled. The relative risk for mortality was calculated with a 95% confidence interval using the fixed effects model. A relative risk of more than 1 indicates a higher risk of death in the first group named (treatment group). The risk difference for the mortality data was not calculated as the event rates were very different. Data for the length of ICU stay and the length of hospital stay were examined for skewedness to assess whether or not to calculate the weighted mean differences for these data (Altman 1996). Subsequently, the weighted mean difference was calculated for the length of ICU stay but not for the length of hospitalisation.

Results

Description of studies

For more detailed description of individual studies, see table of included studies.

Our search found two randomised controlled trials meeting the inclusion criteria, which together involved 1202 patients (data available for 1075 patients). Both trials included data regarding mortality, length of hospitalisation and length of ICU stay.

Risk of bias in included studies

Using predefined criteria (Schulz 1995), both trials were judged to have inadequate allocation concealment (C) as their method of randomisation was alternation.

Effects of interventions

Two randomised controlled trials met the study inclusion criteria. Data regarding overall mortality, duration of hospital stay and of ICU stay were available in both trials.

In the trial by Mattox 1989, 89 patients were excluded due to protocol violations and 38 excluded for other reasons (that is, transfer of patient to another hospital, incomplete or irretrievable hospital records). The author was contacted in order to obtain the data required for an intention‐to‐treat analysis but the relevant data were no longer available. Consequently data are available for only 784 of the 911 randomised patients. 
 
 In the trial by Chang 1995, 43 patients were excluded due to protocol violations. The trial author was contacted and data required for an intention‐to‐treat analysis were obtained. As a result, data are available for all 291 patients randomised.

The results were therefore based on 1075 patients from the two trials.

Mortality

The pooled relative risk of mortality in these patients was 1.13 (95% CI 0.97 to 1.32).

Length of ICU stay

The weighted mean difference of the length of ICU stay was 1.66 (95% CI 0.33 to 2.98).

Length of hospital stay

The weighted mean difference of the length of hospitalisation was not calculated as the standard deviation was very large in comparison to the mean, indicating skewed data. The mean length of hospitalisation in the MAST group was ten days in the Mattox 1989 study and 11 days in the Chang 1995 study, while in the comparison group the mean length of stay was eight days in the Mattox 1989 study and nine days in the Chang 1995 study.

Discussion

This systematic review summarises evidence from randomised controlled trials assessing the efficacy of MAST/PASG in trauma patients.

No evidence was found to show that mortality, length of hospitalisation or ICU stay were decreased with the application of MAST/PASG in trauma patients (excluding those suffering from extremity fractures). Based on the pooled relative risk for the mortality data, patients randomised to the MAST group are 1.13 times more likely to die than those in the control group (95% CI 0.97 to 1.32). MAST/PASG application may therefore be associated with an increased mortality in trauma patients (excluding extremity fractures).

The data indicate that patients randomised to the MAST group have an increased hospital stay and there is no evidence to suggest that MAST/PASG reduces the length of hospitalisation in trauma patients. However, the skewedness of the data meant that it was inadvisable to calculate the weighted mean difference, therefore conclusions drawn from these data should be cautious. Duration of ICU stay may also be increased in those patients treated with MAST/PASG by about two days. It was anticipated that the possible differences in hospital practices that may have influenced length of hospitalisation and ICU stay were similar for the intervention and the control groups. The effect estimate as opposed to raw data was pooled across studies, therefore a valid pooled estimate of the effect of MAST was estimated even if the hospitals in the two trials had different policies. 
 
 The possibility of a survivor effect occurring must be taken into account. This describes the fact that increased length of hospital stay and/or length of ICU stay may be due to the increased survival of sick patients with treatment. However, in this case, as mortality in the treatment group was higher, this explanation is unlikely.

Mortality, length of hospitalisation and ICU stay may all be increased with the application of MAST/PASG in trauma patients (excluding those suffering from extremity fractures).

In 1977, the Committee on Trauma of The American College of Surgeons added antishock trousers to the list of essential equipment for ambulances (Schwab 1983). This equipment is expensive (Schneider 1989), and its use seems to have been advocated in the absence of adequate evidence. The results of this review do not support the use of MAST/PASG for the treatment of trauma patients (excluding those suffering extremity fractures).

It is possible that the concept of MAST itself (that is, autotransfusion from the lower limb) may be ineffective in the treatment of traumatic shock. If this were true then the procedure of raising the legs of a patient who is hypovolaemic due to traumatic shock must also be reassessed.

Both trials have poor quality allocation concealment; however, as these were the only trials identified by our searches, they are included and the possibility of bias must be taken into consideration.

Although the losses from the trial involving 911 patients (Mattox 1989) have been explained, the groups remaining are very unbalanced (MAST n=345, no‐MAST n=439). This suggests that the alternate allocation was violated or that most of the losses occurred from one group. This hypothesis identifies a possible source of bias in the review. The fact that an intention‐to‐treat analysis was unable to be performed on these data means that the results should be interpreted with caution.

Authors' conclusions

Implications for practice.

The results of the trial by Mattox 1989 led to the discontinuation of MAST use in ambulances of the Houston Fire Department where their presence had previously been mandatory. This systematic review does not support the use of MAST/PASG in trauma patients (excluding extremity fracture). However, it should be recognised that, due to the poor quality of the trials, conclusions should be drawn with caution.

Implications for research.

The results of this systematic review suggest that MAST/PASG may increase the mortality, length of hospitalisation and ICU stay in trauma patients (excluding extremity fracture). It would be useful to determine, with large randomised controlled trials, whether autotransfusion in general is ineffective in the treatment of hypovolaemic patients due to traumatic shock.

Methodological qualities of subsequent trials could be improved by including data from protocol violations; this would enable intention to treat analyses to be conducted, making the conclusions more reliable.

What's new

Date	Event	Description
6 August 2009	Review declared as stable	This intervention is no longer in use.

History

Protocol first published: Issue 4, 1999
 Review first published: Issue 4, 1999

Date	Event	Description
11 July 2008	Amended	Converted to new review format.
5 April 2007	New search has been performed	The searches were last updated in April 2007, no new trials for inclusion were identified.

Appendices

Appendix 1. Search strategy

MEDLINE (Ovid) to April 3 2007 
 1. (MAST adj3 suit$).mp. 
 2. PASG.mp. 
 3. (G‐suit$ or Anti‐G suit$).mp. 
 4. ((shock or antishock or anti‐shock or pneumatic or pressur$) adj3 (suit$ or trouser$ or garment$)).mp. 
 5. exp Gravity Suits/ 
 6. or/1‐5

EMBASE (Ovid) to April 3 2007 
 1. (MAST adj3 suit$).mp. 
 2. PASG.mp. 
 3. (G‐suit$ or Anti‐G suit$).mp. 
 4. ((shock or antishock or anti‐shock or pneumatic or pressur$) adj3 (suit$ or trouser$ or garment$)).mp. 
 5. exp Mast Suit/ 
 6. or/1‐5

Central to 2007, Issue 1 
 #1 MAST near3 suit* 
 #2 (shock or antishock or anti‐shock or pneumatic or pressur*) near3 (suit* or trouser* or garment*) 
 #3 PASG 
 #4 G‐suit* or Anti‐G suit* 
 #5 #1 or #2 or #3 or #4

NRR to 2007, Issue 1 
 # 1MAST and suit* 
 #2 (shock or antishock or anti‐shock or pneumatic or pressur*) and (suit* or trouser* or garment*) 
 #3 PASG 
 #4 G‐suit* or Anti‐G suit* 
 #5 #1 or #2 or #3 or #4

PubMed searched on April 3 2007 (last 6 months) 
 #1 MAST near3 suit* 
 #2 (shock or antishock or anti‐shock or pneumatic or pressur*) near3 (suit* or trouser* or garment*) 
 #3 PASG 
 #4 G‐suit* or Anti‐G suit* 
 #5 #1 or #2 or #3 or #4 
 #6 "Wounds and Injuries"[MeSH] 
 #7 injur* or accident* or trauma* or wound* or contusion* or rupture* or lacerat* or stab or stabbed or stabbing* 
 #8 #6 OR #7 
 #9 #5 AND #8 
 #10 randomized controlled trial OR randomly 
 #11 #9 AND #10

ISI Web of Science searched 4 April 2007 
 1. MAST and suit* 
 2. antishock and trouser* 
 3. anti‐shock and trouser* 
 4. antishock and garment* 
 5. anti‐shock and garment* 
 6. PASG 
 7. G‐suit* or Anti‐G suit* 
 8. 1 or 2 or 3 or 4 or 5 or 6 or 7

Data and analyses

Comparison 1. MAST/PASG versus no‐MAST/no‐PASG.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 death	2	1075	Risk Ratio (M‐H, Fixed, 95% CI)	1.13 [0.97, 1.32]
2 length of hospitalisation	2	1075	Mean Difference (IV, Fixed, 95% CI)	1.80 [‐0.84, 4.44]
3 length of ICU stay	2	1075	Mean Difference (IV, Fixed, 95% CI)	1.65 [0.33, 2.98]

1.1. Analysis.

Comparison 1 MAST/PASG versus no‐MAST/no‐PASG, Outcome 1 death.

1.2. Analysis.

Comparison 1 MAST/PASG versus no‐MAST/no‐PASG, Outcome 2 length of hospitalisation.

1.3. Analysis.

Comparison 1 MAST/PASG versus no‐MAST/no‐PASG, Outcome 3 length of ICU stay.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Chang 1995.

Methods	Prospective randomised study; randomisation by alternate day assignment.
Participants	291 traumatic shock patients. These were adults >15 years, with blunt or penetrating injuries and a systolic blood pressure of 90 mmHg or less with clinical signs of hypotension, including altered level of consciousness and evidence of decreased vascular perfusion as determined by EMS in the field.
Interventions	1. PASG 2. No‐PASG
Outcomes	overall mortality (including those patients dead on arrival (DOA) and those dying in the trauma unit (DER)), number of patients dead on arrival, length of hospital stay, length of Intensive Care Unit stay, number of patients intubated, number of patients with chest tubes, number of patients with thoracotomies.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	C ‐ Inadequate

Mattox 1989.

Methods	Prospective randomised study; randomisation by alternate day assignment.
Participants	911 patients included in the study. These patients were adult >15 years and victims of blunt or penetrating injuries. They presented with: 1. systolic blood pressure of 90 mmHg or less at the time of initial prehospital assessment by paramedics from the City of Houston EMS. 2. were transported by ambulance to the Ben Taub General Hospital. Patients were excluded from the study for the following reasons: pregnancy evisceration impaled objects in a body region that would be encompassed by the pneumatic garment.
Interventions	1. MAST 2. No‐MAST
Outcomes	overall mortality (including patients dead on arrival (DOA) and those dying in the hospital itself), number of patients DOA, length of hospital stay, length of ICU stay.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	High risk	C ‐ Inadequate

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Barriot 1988	Trial involves the comparison of high and low inflation pressures of MAST in severe haemorrhagic shock.
Bickell 1985	Study involves the effect of MAST on the trauma score which is assessed at presentation to the hospital. This is not an outcome that we are interested in.
Bickell 1987	Data included in the prospective MAST study in 911 patients, Mattox et al. 1989.
Mattox 1986	Data included in the prospective MAST study in 911 patients, Mattox et al. 1989.
Pepe 1986	Summary paper of data from the trials reviewed.

Contributions of authors

KD screened the citations for eligibility, contacted authors, extracted data, entered data into RevMan and wrote up the review. IR helped to develop the protocol, screened citations, extracted data, contacted authors and helped to write the review.

Declarations of interest

None known.

Stable (no update expected for reasons given in 'What's new')