Finding the key to improving survival The Holy Grail of Trauma.

Finding the key to improving survivalThe Holy Grail of Trauma

Publication Explosion Trauma and Blood Transfusion – 1990 - 2010

Improving survivability in traumaOutline

Magnitude of the improvement gap

Summary

The evidence from these options

Transfusion options

Goal-directed therapy

11

22

33

44

55

Evidence for these options

http://www.llf.org.uk/resource/llf_blood_transfusion_jpeg.jpg

• Chart and autopsy studies

• Before and after implementation of trauma systems

• Time to death

• Country comparisons

How do we quantify the potential improvement gap?

Preventable DeathsEsposito TJ, et al. J Trauma. 1995 Nov;39(5):955-62

41 preventable deaths inMontana in 1990

Preventable DeathsEsposito TJ, et al. J Trauma 2003; 54: 663-70

25 preventable deaths inMontana in 1998

• Preventable Deaths at Royal London Hospital before & after implementation of a trauma system

Preventable DeathsDavenport RA, et al. B J Surg 2010; 97: 109-117

Year PreventableDeaths (P/P)

Percent

2004 6 0.87

2005 2 0.24

2006 2 0.19

All 10 0.38

• Mortality outcomes in 18 trauma hospitals and 51 non-trauma centers in 14 states (18,198 patients)– Mortality rates – Death in hospital (adjusted):

• 7.6% trauma centers• 9.5% non-trauma• Relative risk 0.80 (0.66-0.98)

– Subgroup analysis suggested benefit restricted to those under 55 years of age (or no improvement in the ‘elderly’)

Trauma SystemsMacKenzie EJ, et al. N Engl J Med 2006; 354: 366-78

1.9%

• Trauma patients treated in the Netherlands before and after implementation of a trauma system (n=68,041)– Mortality rates – Death in hospital:

• 2.6% before• 2.3% after• Relative risk 0.89 (0.80-0.98)

– Adjusted OR – 0.84 (0.76-0.94)– Subgroup analysis – trauma system appeared to

assist the less severely injured and elderly patients

Survivability in TraumaTwijnstra MJ, et al. Ann Surg 2010; 251: 339-43.

0.3%

When do patients die in the first 24 hrs?Acousta JA, et al J Am Coll Surg 1998; 186: 528-533

35% died in the 1st 15 min

Resuscitation Outcomes ConsortiumMinei JP, et al. Ann Surg 2010; 252:149-157

Nothing at the scene has changed in 30 years

Baker CC, et al. Am J Surg 1980; 140:144-150

53%

10 of 437 (2.3%) deaths deemed preventable

Country comparisonsChristensen MC et al. J Trauma 2010; 69: 344-352

0.3-6% of all trauma deaths might be preventable(and a subgroup of these might be salvageable with transfusion strategies)

Bottom line?

• 36.8 million ‘injuries’ in 2009– 122/1000 population

• 182,479 deaths– 60.5/100,000

• If 1% preventable…– And half of the deaths ‘transfusion-related’…– 913 deaths might be preventable with better

transfusion support

Incidence of Injury in USA2009 CDC Statistics

Prevention – Seat BeltsThoma T. Ann Emerg Med 2009; 54: 837-839

Prevention – Gun controlPeople don’t kill people, guns kill people

22,523 gun licenses refused or revoked between 1999 and 2008 from individuals deemed a potential risk to themselves or to others

Prevention – Gun controlGuns don’t kill people, people kill people

1 in 100 U.S. Adults Behind Bars, New Study SaysNY Times Feb 28, 2008



Summary


Transfusion options

Goal-directed therapy (my option)

11

22

33

44

55



✓

Transfusion is not the only optionNon-transfusion Options that work

33 The implementation of organized trauma systems translates into a 1.9% absolute increase in survivability

Trauma Systems

RCT published in the NEJM in 1994 showed a 8% absolute increase in survival for penetrating traumas

Scoop and Run

11

Insurance at the time of gunshot injury translates into a 3% absolute decrease in mortality

Insurance

2

Bickell et al. NEJM 1994; 331: 1105-9Dozier KC, et al. J Am Coll Surg 2010; 210: 280-5.MacKenzie et al. N Engl J Med 2006; 354: 366-78

“We propose that the administration of exogenous estrogen will improve outcomes after injury and hemorrhage”

“Imagine the …possible benefits to early estrogen administration in a testosterone-laden biker, screaming obscenities in your trauma bay.”

Women RuleBullard MK et al, Surgery. 2010 Jan;147(1):134-7.

http://texasbestgrok.mu.nu/images/Elastigirl.jpg

Non-transfusion options that don’t work

Resident working hoursResident working hours

Restriction of resident working hours did not improve outcomesRestriction of resident working hours did not improve outcomes33

A study from Johns Hopkins Hospital found no improvement in survival when they compared novice surgeons to experts

The senior surgeon

11

A study from 10 trauma hospitals in North America found no improvement in survival with faster transport times

Transport speed

2

Haut ER, et al. Arch Surg 2009; 144: 759-64Newgard CD, et al. Ann Emerg Med 2010; 55:235-46Helling TS, et al. J Trauma. 2010: 69:607-12

What should you put in the box?Transfusion Options

WW

Recombinant factor VIIa

Fresh blood

Formula resuscitation

11

22

33

44

55

66

Fresh warm whole blood

http://ecx.images-amazon.com/images/I/31H6wFfIl7L._SL500_AA300_.jpg

ABO identical plasma

Fibrinogen concentrates

KD Boffard, B.Riou, B.Warren et al. J.Trauma. 2005; 59:8-18

1st RCT rVIIa• Conclusion: Recombinant FVIIa resulted

in a significant reduction in RBC transfusion in severe blunt trauma. Similar trends were observed in penetrating trauma. The safety of rFVIIa was established in these trauma populations within the investigated dose range.

Boffard KD, et al. J Trauma. 2005;59:8-15

Resident guidance suggestedRESIDENT GUIDANCE RECOMMENDEDRESIDENT GUIDANCE RECOMMENDED

FOR WACKY STATISTICAL METHODS FOR WACKY STATISTICAL METHODS

Transfusion(units RBC)

rFVIIa

0 6

RandomisationArrivalat ERTrauma

Placebo

2 studies (Blunt & Penetrating)

Trial Design

200 ug/kg 0100 ug/kg 1100 ug/kg 38

n=301

Boffard KD, et al. J Trauma. 2005;59:8-15

* p=0.07

RBC Control

RBC R7a

Blunt 7.2 u 7.8 u

Penetrating 4.8 u 4.0 u

No Effect on Transfusion RateBoffard KD, et al. J Trauma. 2005;59:8-15

Exclude patients who bled to death in the 1st 2 days?

Survival (not ITT)Boffard KD, et al. J Trauma. 2005;59:8-15

• Prospective, randomized, double-blinded, multicenter trial (150 hospitals in 26 countries)– 3 doses r7a 200/100/100 ug/kg - $30K– Up to age 70– Still bleeding with shock/hypotension/acidosis

after 4 units RBC

Recombiant Factor VIIa - CONTROLHauser CJ, et al. J Trauma 2010; 69: 489-500

• Powered to detect a 16.7% mortality reduction assuming a 30% baseline mortality

• Planned interim analysis

• Stopped early due to high likelihood of futility

• 573 enrolled, 560 dosed, 554 in ITT

• No difference in mortality (11% vs 11%)

Recombinant Factor VIIa - CONTROLHauser CJ, et al. J Trauma 2010; 69: 489-500

• Does r7a prevents FFP transfusion which prevents RBC transfusions?

Recombinant Factor VIIa - CONTROLHauser CJ, et al. J Trauma 2010; 69: 489-500

• Arterial TE events were more common in r7a treated patients OR 1.68 (1.2-2.4, p=0.003)– Risk attributed to patients over 65 years– 65-74 yrs – OR 2.12 (0.95-4.71, p=0.07)– >75 yrs – OR 3.02 (1.22-7.48, p=0.02)

Safety Profile n= 4119Levi M, et al. NEJM 2010; 363: 1791-1800.

What about formula driven resuscitation or 1:1:1 or DCR?

• >10 units of blood (RBC/whole blood) in 24 hrs– This type of trial design may excluded severely injured patients who died

before #10

• Divided patients into 3 groups based on similar ratios of FFP:RBC and mortality

• Analyzed to determine the effect of the FFP:RBC ratio on mortality

• Transfusion protocol not disclosed• Between 2003-05 – 5,293 admitted of whom 246 (4.6%) were

included• 95% penetrating trauma

Iraq 2003-2005Borgman et al. J Trauma 2007; 63: 805-813.

Variable Not muchN=31

MediumN=53

Lots (1:1.4)N=162

Thorax injury 26% 9% 7%

Hemoglobin 94 108 109

Base deficit 13 9 8

Heart rate 122 118 111

sBP 90 98 97

Crystalloid 1.8 L/hr 0.6 0.5

RBC 4 units/hr 0.9 0.8

Plasma 2 6 12

Plasma 0.1 unit/hr 0.3 0.6

Factor 7a 16% 26% 36%

Variable Not much Medium Lots

n 31 53 162

Mortality 65% 34% 19%

Median time to death

2 hrs (1-4) 4 hrs (2-16) 1.6 days (4 hours – 6.5 days)

Number of deaths

20 18 31

OutcomesBorgman et al. J Trauma 2007; 63: 805-813.

• High FFP:RBC ratio results in a 55% absolute risk reduction in mortality!– High FFP:RBC ratio decreased the hourly transfusion

rate

Their conclusionBorgman et al. J Trauma 2007; 63: 805-813.

“If you want to get people to believe

something really, really stupid, just stick a

number on it.”Author Charles Seife

Proofiness: The Dark Arts of Mathematical Deception

“If you want to get people to believe

something really, really stupid, just stick a

number on it.”Author Charles Seife

Proofiness: The Dark Arts of Mathematical Deception

QUOTABLEQUOTABLE

Study n Highest FFP Lowest FFPDuchesne 135 26% 88%Maegele 713 24% 46%Holcomb 466 40% 60%Scalea 250 No differenceKashuk 133 8% 40%Sperry 415 28% 35%Teixeira 383 26% 90%Zink 466 26% 55%

Other Retros

Median 26% 55% Delta 29%!

Of course, same goes for plateletsInaba K, et al. J Am Coll Surg 2010; 211: 573-9.

Different patients:

Different outcomes:

• 2 way analysis:(1)the effect of the ratio at 24 hours on outcome

(2)the effect of the ratio on outcome in a time-dependent analysis

• Median time to the first RBC and first FFP was 18 and 93 minutes, respectively

• The start times for the first FFP ranged from 24 to 350 minutes!

Snyder – Confirms survivorship bias J Trauma 66:358-364, 2009

Snyder - AlabamaJ Trauma 66:358-364, 2009

0.37;0.22-0.63 0.84, 0.47-1.50

• Ruptured AAAs between 1987 and 2007 (note: pre 1:1 paper by Borgman)

• Includes 128 patients transfused >10 units• Transfusion at the discretion of the MDs• Volume of FFP did not impact survivability• In multivariate analysis, ratio of FFP (<1:2)

resulted in a 4-fold increased risk of death• Only possible mathematical conclusion –

more RBCs = more deaths

Indication CreepMell MW, et al. Surg 2010; 148: 955-62.

http://www.learningradiology.com/caseofweek/caseoftheweekpix/aneurysmrupture2.jpg

FFPRBC

• Another way to look for effects on outcomes

• Unfortunately:– Selection bias results in different patients

between the two groups– Not the same time period – other changes to

care have occurred

The pre- and post-MTP studies

• They were able to change transfusion practice

Military Before, AfterSimmons JW, et al. J Trauma 2010; 69: S75-80.

• Formula-driven resuscitation was associated with an increased risk of MT despite no differences in baseline characteristics


• They successfully managed patients “better”– Warmer on arrival (96.5 to 98.2°F)– Less crystalloid exposure in first 12 hours (14 vs. 9 L)– More FFP (8 to 14 U)– More platelets (1 to 2 U)– “Better” ratio (0.54 to 0.76)– Faster transport– CAT-tourniquet for every soldier– New medic resuscitation guidelines


Miltary Before, After (n=777)Simmons JW, et al. J Trauma 2010; 69: S75-80.

P=0.12

TRFL – Pilot Feasibility Study

35 patients - 16 months

Lab = 14 patients

32 patients

1:1:1 in 75%ratio 1.2:1:117% death

Lab q2h in 100%ratio 2:1:0.614% death (24h)

3 excluded

1:1:1 = 18 patients

• Retrospective cohort 2004-07 of trauma pts admitted to ICU, civilian

• Patients transfused ≥5 RBC units during admission– Compared: <27 days vs. >28 days

• Matched by RBC Tx (+/- 1 unit)• Primary outcomes were deep vein thrombosis

and in-hospital mortality (?)

Fresher Blood in TraumaSpinella et al. Crit Care. 2009;13(5):R151.

• 270 patients identified of whom only 202 (75%) could be matched– Patients receiving ‘older’ blood were more

likely to have blunt injury (96 vs. 89%)

• RBC storage age:– Maximum -19 vs. 34 days– Median - 14 vs. 20 days

Fresher Blood in TraumaSpinella et al. Crit Care. 2009;13(5):R151. Epub 2009 Sep 22.


Survived (%)(n=161)

Died (%)(n=41)

78%(56/72)

22%(16/72)

79%(30/38)

21%(8/38)

83%(10/12)

17%(2/12)

81%(65/80)

19%(15/80)


12.8%


• 430 transfused trauma patients over 7 years

• ‘amount’ of young vs old (>14 days) RBC comparison

Young vs. OldWeinberg JA, et al. J Trauma 2008; 65: 794-798.

• Reduce storage lesion• Improves cardiac output• Improves

microcirculatory hemodynamics

• Improves O2 consumption

• Already in ‘right’ ratio

Why might WFWB be better?Spinella PC, et al. J Trauma 2009; 66: S69-76.

Spinella PC. Crit Care Med 2008; 36: S340-45.

• Warmer• More concentrated• More ‘functional’• Less additives• Less anticoagulants• Corrects

coagulopathy more efficiently

• Decreases mortality

Warm fresh whole blood

• Damage control resuscitation using warm fresh whole blood: a paramount role for leukocytes and derived microparticles in the prevention of coagulation abnormalites

• Transfusing functional white blood cells is one of the singularities of WFWB!!

Warm Fresh Whole BloodMorel N, et al. J Trauma 2010; 68: 1266-7.

http://en.wikipedia.org/wiki/File:The_Scream.jpg

• Jan 2004-October 2007, retrospective– WFWB + RBC, FFP (but not PLT)– vs. components only

• Outcome – 24 hr and 30 day mortality

• Well matched – except WFWB patients were colder on arrival (earlier cohort)

• Use of WFWB decreased over time – why?

Warm Fresh Whole BloodSpinella PC et al. J Trauma 2009; 66 (4 Suppl): S69-S76

• Missing whole blood unit recipients– This paper: 100 patients x 5 Units (median)

= about 500 units

• 2008 review by same author quotes more than 6000 units transfused

• Who got the other 5500 units?


Spinella PC. Crit Care Med 2008; 36: S340-45


Key words in the discussion on the limitations of the study…

“…because of the time required to initiate and collect WFWB…”

Fresh Warm Whole BloodSpinella PC et al. J Trauma 2009; 66 (4 Suppl): S69-S76

• Retrospective analysis of transfused trauma patients, 2000-2008– Between 2000 and 2008, the use of ABO-non-

identical plasma increased 200%– Propensity matching required

• 10 units of group O plasma, 8 units of group A plasma, and 8 units of group B plasma thawed at all times

ABO-identical plasma saves livesInaba K, et al. Arch Surg 2010; 145: 899-906


• Why would any group O patient ever get non-identical plasma at their center?1) Early in the time period there was no thawed

plasma and rapidly bleeding patients were transfused a few units of AB plasma until the blood group was processed (sicker & earlier time period)

2) These patients were bleeding so fast that they depleted the group O inventory (sicker)


Whatever it was, it was not random!


ABO-non-identical plasma harms 1 in 8!

Fibrinogen:RBC RatioStinger HK, et al. J Trauma 2008; 64: S79-85

Same patients as Borgman paper…same number…same time…same place

Fibrinogen ConcentratesThe next trend?

16 grams

Acta Anaesthesiol Scand. 2010;54:111-7. Epub 2009 Oct 26Anaesthesia. 2010;65:199-203. Epub 2009 Nov 30Scand J Clin Lab Invest. 2010;70:453-7

9 grams

12 grams

• Retrospective analysis of trauma patients transfused >5 u/24 hours

• They use ROTEM to decide what to give– Increased FIBTEM MCF – 2-4 g fibrinogen– Increased EXTEM MCF– Platelets– Increased EXTEM CT – PCC 1000-1500 IU– When do they give FFP?

Fibrinogen as per FIBTEMSchochl H et al. Critical Care 2010; 14: R55

http://en.wikipedia.org/wiki/File:Flag_of_Austria_(state).svg

• N=149 patients over 4 years RBC>5/24– Excluded 15 that died in <60 min and 3 that got

nothing but RBC– Severely injured – mean ISS 38– Median 10 RBC/24 hours– Only 3/131 did NOT get fibrinogen concentrates!

(median 7 g/24 hours) – 0.8g:RBC– 30 treated with PCC, 21 FFP, 29 platelets!– Predicted mortality 34%, observed 24%

Fibrinogen as per FIBTEMSchochl H et al. Critical Care 2010; 14: R55

Theusinger OM et al. Curr Opin Anesth 2009; 22: 305-12.



Summary


Transfusion options

Goal-directed therapy

11

22

33

44

55



✓✓✓

Goal Directed TherapyThe individualized approach

E=mc2

No bleeding =No components

CBC, INR, fibrinogen q1h

E=mc2

AA

B

C

CBC, INR, fibrinogen q1h

RBCs <10 and surgicalcontrol planned

ONLY RBCS

CBC, INR, fibrinogen q1hLAB DRIVEN

• ABC Score >2 = 85% chance of MT

Predicting who will need massive transfusionCotton BA, et al. J Trauma 2010; 69: S33-9.

✓

✓✓

X

3

TASH-score updatedMaegele M, et al. Vox Sang 2010; Aug 24: epub

• Independent predictor of MT1. 27.8% of patients receiving unmatched went

on to receive 10 U or more in the first 6 hours (vs. 5.3% of those that did not)

2. 29.3% of patients receiving unmatched went on to receive 10 U or more in the first 12 hours (vs. 1.8% of those that did not)

3. My hospital – 26% vs. 1.8%

Unmatched RBC predictsNunez TC, et al. Transfusion 2010; 50: 1914-20

Inaba K et al. J Trauma 2008; 65: 1222-6

• 1,685 trauma patients transfused <10 RBC– 30.6% received plasma in first 12 hours– Half had an ISS>25

• 284 matched pairs identified– 2.9 U RBC over first 12 hours– 3.0 U FFP for the plasma group– Groups were well matched

Harm to patients who do not require massive transfusion

Inaba K, et al. J Am Coll Surg 2010; 210: 957-65.

Collateral DamageHarm to patients who do not require massive transfusion

Inaba K, et al. J Am Coll Surg 2010; 210: 957-65.

Number Needed to Harm = 12

“Until more data are available, caution should be exercised in using fixed ratios of blood components for all except early resuscitation of the most severe trauma cases as all blood products carry risk that may outweigh therapeutic benefit if used in excess”

My favorite quoteMorley SL. Arch Dis Child Educ Pract Ed 2010; epub

“Such strategies should also be regarded as ‘resuscitation’ in the most acute sense and as soon as hemorrhage is controlled and the patient’s clinical status has stabilised, then titration of products based on blood testing should be re-instituted to reduce the risk of overtransfusion.”

My favorite quoteMorley SL. Arch Dis Child Educ Pract Ed 2010; epub

• 0.3-6% of all trauma deaths may be preventable (some with transfusion)– We might be at the end of the journey for improving

survival in trauma

• No clear successes to date: rVIIa, 1:1:1, fresh blood, WFWB, ABO-identical or fibrinogen concentrates

• Stick with an individualized goal-directed approach until RCTs show us a better way

Summary

Lots of great ideas, lots of hype,

but no clear winners in this quest

Finding the key to improving survival The Holy Grail of Trauma.

Documents