TRALI

Cases and Chapter Session

Tanya Petraszko MD FRCPC

CBS BC & Yukon
January 2010
What is Transfusion-related
Acute Lung Injury?
 Sudden onset of “Acute Lung Injury” occurring
within 6 hours of a transfusion
 Acute Lung Injury
 Hypoxemia
 New bilateral chest X-ray infiltrates
 No evidence of volume overload

Canadian Consensus Conference Definition. Kleinman et al. Transfusion 2004;44:1774-89

 Features of TRALI

 SOB and abnormal CXR

 Fever, hypotension

 onset is usually within the first 1-2 hours of

transfusion, but may be delayed up to 6
hours
Normal CXR Patient’s CXR
 TRALI Mortality
 TRALI appears to be the most common cause of
transfusion-related mortality

 Reports to the FDA of transfusion-related

deaths:
 TRALI > bacterial sepsis > acute hemolysis

 Estimated rate of fatalities is 5 – 10%

Transfusion Related Fatalities
Reported to FDA 2003-2005

Cause #cases (% of fatalities)

 TRALI 72 (32%)
 Bacterial sepsis 25 (11%)
 AHTR 23 (10%)
 Other 71 (32%)
 Not clearly tx related 34 (15%)
 Total 225 (100%)
 Leading Cause of Adverse Transfusion
Events reported to Health Canada

133
Number of Adverse Reactions

140
118 113
120 112
95
100
80 Total AR
60 45 48 46 TRALI
42
40 30
17
20 10
0
2001 2002 2003 2004 2005 2006
 Evolution of Transfusion Risks
HIV and HCV vs TRALI
1 / 102
HIV
Transfusion Risk per Unit

HCV
1 / 103
Estimated TRALI risk

1 / 104

Estimated TRALI fatality (?)

1 / 105

1 / 106

1984 1988 1992 1996 2000 2004

Adapted from Goodnough L and Aubuchon J
 WBC Antibody Hypothesis
 Antigen-Antibody reaction triggers TRALI
Release of substances
causing pulmonary
endothelial damage
Activation and capillary leak
Donor
-Anti-HNA Lodged in
-Anti-HLA I pulmonary
-Anti-HLA II capillaries

Recipient WBC
- Neutrophils
- Lymphocytes
- Monocytes

Courtesy of Y. Lin
Clinical Correlation
 Right, single lung transplant

 10 weeks after transplant,

transfused PRBCs for anemia

 HLA B44 antibody in PRBC

donor

 HLA B44 antigen present in

lung donor tissue only

 Elegant clinical example of

targeted endothelial injury
from a blood transfusion

Dykes et al. Br J Haematol 2000.

What evidence is there for this
 Many studies demonstrate HLA or
leukoagglutinating alloantibodies in
donors
 Several examples of donor-recipient
antigen/antibody concordance
 Anti HLA class I and II implicated
 Anti-neutrophil antibodies implicated
Evidence for antibody theory
 ex vivo isolated rabbit lung model
 severe lung vascular leakage was
reproduced using a human anti-neutrophil
antibody in the presence of cognate,
human 5b-positive neutrophils
 In contrast, no vascular leakage was noted
in lungs perfused in the absence of either
 antibody,
 neutrophils, or a
 complement source.
 No permeability increase occurred with the
use of 5b-negative neutrophils

Seeger Blood. 76:1990.

Antibody theory animal models
 isolated perfused rat lungs
 perfused them with plasma containing anti HNA-2a mAb
and human neutrophils.
 If neutrophil expression of HNA-2a antigen was >70%, ALI
was manifest;
 if antigen expression was <30%, ALI was not manifest.
 However, if the lungs were primed with fMLP then lung
injury was induced in the group with < 30% antigen
expression.
 Thus, if there is no relevant first hit (priming with fMLP),
TRALI may not be manifest even with cognate antibody
unless the cognate antigen expression meets a certain
threshold.
 also demonstrated that anti-HNA 2a directly activates
neutrophils in the absence of complement

Sachs Blood. 107(3): 2006.

 Anti-neutrophil Antibodies (HNA-2a)

+FMLP

Slide from Looney MD, UofC, SF Sachs et al, Blood 2006.

Non antibody mediated TRALI
 15% of typical TRALI cases fail to
demonstrate antibody
 alternate hypothesis that some cases
may be non-antibody mediated
 Biologically active molecules
contained in stored blood products
 lysophosphatidylcholines
 interleukins
What evidence is there for this
 Rats pre-treated with LPS developed acute lung
injury (ALI) when the lungs were perfused with
plasma
 Rats did not get ALI with LPS alone

 Rats pretreated also got ALI with lipids extracted

from stored 42 day old human RBCs or platelets
 Did NOT get ALI if given lipids from 5 day product
 thus plasma or lipids from fresh, human RBCs or
platelets did not cause lung injury.

Silliman CC J. Clin.Invest 101(7): 1998

Silliman CC Transfusion. 43: 2003
Two Event Model
 Recipients of blood containing anti
WBC antibodies don’t all get TRALI
 perhaps two events required to
manifest TRALI (as in ARDS)
 Predisposing clinical condition
 Infusion of biologically active mediators
 First event results in activation of
pulmonary endothelium with
neutrophil priming
 Second event results in activation of
the neutrophils adherent to
pulmonary endothelium and
subsequent lung injury
 Two event hypothesis
2nd event: Transfusion
of biologically active
1st event: Underlying
lipids or antibodies
clinical condition of
patient (inflammation, Activation
infection, surgery)

Release of substances
causing pulmonary
endothelial damage
and capillary leak

Activation of pulmonary
endothelium with increased
adhesion molecules

Silliman et al. Blood 2003;101:454-62

Kleinman et al. Transfusion 2004;44:1774-89
Evidence from animal models
 2-event in vivo rat model
 first event
 saline (NS) or
 endotoxin (LPS)
 second event
 plasma from stored, human packed red blood cells
(PRBCs) or
 antibodies (OX18 and OX27) against rat MHC Class I
antigens
 NS treated rats did not develop TRALI
 With plasma nor with antibody
 LPS treated rats developed ALI
 With plasma and with antibody

Kelher Blood. 113(9):2009.

 Neutrophil depleted rats had experiment
repeated and all failed to demonstrate ALI
 Confirmed that ALI is neutrophil dependent
 also demonstrated that anti-MHC Class I
localized to the neutrophil surface
 They concluded that TRALI is the result of 2
events with the examples of the second
event being plasma from stored blood and
antibodies that prime neutrophils.
Case of recurrent TRALI
52 year old female 4 months post allo BMT for myelofibrosis;
transfusion dependent

2007-02-10
 received 2 units red cells uneventfully as an outpatient.
 Discharged but became acutely SOB on way home,
returned to ER
 CXR showed bilateral new infiltrates
 Decreased sats, intubated and ventilated for 30 hours
then extubated and recovered.

Results of Investigation
 Recipient results: post- sample: Negative

 Donor 1st unit negative

 Donor 2nd unit positive for anti HLA antibody

 Crossmatch donor 2 with recipient positive

Case of recurrent TRALI
52 year old female 4 months post allo BMT for myelofibrosis;
transfusion dependent

2007-03-19
 Had been discharged home receiving regular outpatient
transfusions 2x/week
 after transfusion of 2 units of RCC developed new onset
SOB, new CXR infiltrates and PO2<60.
 Required mechanical ventilation.

Results of Investigation
 Recipient results: pre-samples: positive anti HLA
 post-sample: positive anti HLA
 No crossmatch done

 Donor 1st unit positive anti HLA antibody

 Donor 2nd unit positive anti HLA antibody
Two event model
 In first TRALI event, implicated donor
associated platelet unit did not cause
TRALI in the recipient it was given to

 That recipient was not primed or did

not have cognate antigen….
Evidence from animal models
 in vivo mouse model
 BALB/c wild-type mice positive for MHC
Class I antigen were injected with MHC
Class I mAb via the jugular vein and
sacrificed at 2 hours.
 Control mice were knock out mice negative
for the antigen
 Injection of MHC Class I mAb produced
severe ALI in wild type mice with 50%
mortality compared to controls negative for
the cognate antigen.
Looney MRJ. Clin. Invest. 116:(2006)
 MHC I antibody produces severe ALI
in mice
140 180 BALB/K + MHC I mAb
BALB/c + Isotype control mAb
** 160 **
Excess lung water (µl)

120 BALB/c + MHC I mAb

140
100
120

EVPE (µl)
80 100

60 80
60
40
40
20 20

0 0

100 PBS or Isotype control mAb

MHC I mAb
Survival (%)

80
60 **
40
20
0
15 30 45 60 75 90 105 120
(Minutes)
Evidence from animal models
• The MHC Class I mAb binds
throughout the body and prominently
in the lung microvasculature.
• The investigators demonstrated that
this antibody also binds to
neutrophils, but does not cause direct
neutrophil activation.

Looney MRJ. Clin. Invest. 116:(2006)

Evidence from animal models
 Mice were also neutrophil depleted with GR-1 antibody
 these mice were protected from lung injury.
 Thus the absence of circulating neutrophils renders the bound
antibody on lung tissue inert and unable to cause vascular
permeability.
 Using a knockout mouse it was further demonstrated that it was
the Fc gamma receptor which was essential in responding to the
antibody challenge.
 The Fc gamma receptor knock out mice were fully protected from
lung injury, however when wild type neutrophils were transfused
back, the lung injury was restored.
 Thus in this model,
 MHC Class I antibody binds to class I antigen on the endothelium
 then presents the Fc portion of the antibody to the neutrophil
 Neutrophil is then activated thru Fc gamma receptor engagement.
 demonstrated a critical role for neutrophil Fc gamma receptors in
mediating ALI.

Looney MRJ. Clin. Invest. 116:(2006)

Summary so far
 Clinical features of TRALI
 Leading cause of TR mortality
 Majority appear to be antibody
mediated
 Animal models support this

So….why don’t we always see TRALI

when antibody and cognate ag are
present?
Effect of environment on priming
 Mice were housed in specific-pathogen free
barrier conditions vs. non-barrier housing
 Repeated TRALI model
 Barrier mice were protected from TRALI
 non barrier mice did develop ALI.
 Barrier mice demonstrated lower levels of
circulating neutrophils
 suggesting that this is the reason why they did
not respond to the second hit since this model is
neutrophil dependent.
 if barrier mice were primed with LPS prior
to receiving anti-MHC Class I mAb, then
they did develop ALI in a dose-dependent
manner. Looney. J Clin Invest. 119(11): 2009
 Anti-neutrophil Antibodies (HNA-2a)

+FMLP

Slide from Looney MD, UofC, SF Sachs et al, Blood 2006.

Role of neutrophils in ALI
 microfluidics model of the lung microcirculation.
 MHC Class I mAb is immunoadsorbed to the artificial
capillary bed, and the unbound antibody is removed
by washing.
 Next, Fc gamma receptor knockout neutrophils or
wild-type neutrophils are perfused.
 video demonstrates that Fc gamma receptor knockout
neutrophils zip right through the capillaries as they
cannot engage the bound antibody
 wild-type neutrophils all get stuck as they engage the
bound antibody through their Fc gamma receptor.
Effect of hematopoetic cells in
priming
 hematopoietic chimeras TLR4 mice, which
are LPS unresponsive
 TLR4 mice reconstituted with TLR4 bone
marrow were protected in the two-event
model of TRALI (LPS + MHC Class I mAb).
 However, when TLR4 mice were
reconstituted with wild-type bone marrow,
lung injury was manifest
 thus demonstrating that hematopoietic cells
are the critical population that must be
primed to produce lung injury.
Looney. J Clin Invest. 119(11): 2009
Conclusions from animal models
 Well developed animal models support the
two event hypothesis of TRALI.
 effective ‘second hits’ in TRALI
 Neutrophil or
 MHC Class I antibodies or
 plasma or lipids from stored, human RBCs
 Priming of hematopoietic cells is necessary
in experimental TRALI and may be the ‘first
hit’.
 Common to all experimental models,
neutrophils are the key effector cell in
TRALI.
How can we prevent TRALI?
 We have evidence that antibodies are
responsible for TRALI in a primed
host

 How can we identify a primed host?

 Fresher product for the primed host?
How can we prevent TRALI?
 TRALI reduction strategies have focused on
donors with antibodies and specifically plasma
containing products

 Secondary – Measures taken to identify the cause

of a reaction that has occurred and prevent it from
happening again

 Primary – Preventive measures taken to eliminate

the risk before it happens in the first place
How can we prevent TRALI
 Secondary…defer implicated donors
 Donors associated with a reported TRALI are
investigated for anti HLA and anti neutrophil
antibody and deferred if positive
 Pilot 2001 – 2006, SOP 2006
 Primary…avoid donors with anti HLA or anti
neutrophil antibody
 what have blood suppliers done and what
has the effect been
 Donor loss
 Reduction of TRALI
Which donors develop WBC antibodies?

 WBC alloimmunization may occur following

previous exposure to WBCs through pregnancy
or transfusion
 332 female plateletpheresis donors
 17% had detectable anti-HLA antibody
 Frequency of HLA antibodies increased with
pregnancy:
0 pregnancies: 7.8%
1-2 pregnancies: 14.6%
3 or more: 26.3%

Densmore et al. Transfusion 1999;39:103-6

Primary Prevention
 2003 – UK changed component production
policy moving to predominantly male plasma
for transfusion
Chapman et al. Vox Sang 2006;91(Suppl 3):227

2003 2004 2005

TRALI 36 23 23

Highly 22 13 6
likely/probable
SHOT data
 10 years hemovigilance in UK and impact of
preferential use of male donor plasma
 Risk of highly likely/probable TRALI due to
FFP decreased
 15.5 per million units 1999-2004
 3.2 per million units issued 2005-2006 (p=
0.0079)
 Risk of highly likely/probable TRALI due to
platelets decreased
 14 per million to 5.8 per million

 Chapman.Transfusion 2009;49:440-452
TRALI reduction measures CBS-
predominantly male plasma
 CBS moved to predominantly male
plasma for transfusion Oct 2008
 Female plasma diverted fractionation
 TRALI the CBS Experience – Definite and Possible

SMH: GIFT, LIFT, LCT, X-Match when possible CBS:Luminex HLA

Year 2001 2002 2003 2004 2005 2006 2007 2008 2009
Q1
Total 6 6 6 16 14 21 35 19* 1

Tor 4 6 6 10 6 10 26 5 1
Edm 1 0 0 3 3 1 0 0* 0
TRALI Reduction Measures –
Platelet Apheresis

 On July 20, 2009 CBS started asking

all female platelet donors if they’d ever
been pregnant
 Any responding positively redirected to
whole blood
Donor Loss YTD
 June 2009 there were 2394 female
apheresis platelet donors
 By Nov 2009 only 300 active female
apheresis platelet donors
 32% of the remaining donors
converted to whole blood or plasma
(approx 765 donors)
 Net loss of approx 1329 donors
 BC&Y Platelet Imports and Issues (Doses)

20.0 K 25.0%
18.0 K 22.5%
16.0 K 20.0%
Imports and Issues

14.0 K 17.5%

% Imports / Issue
12.0 K 15.0%
10.0 K 12.5%
8.0 K 10.0%
6.0 K 7.5%
4.0 K 5.0%
2.0 K 2.5%
0.0 K 0.0%
2004/05 2005/06 2006/07 2007/08 2008/09
Imports 650 1,335 2,344 3,642 2,559
Issues 12,829 14,318 15,599 17,412 17,719
% Imports/Issue 5% 9% 15% 21% 14%

F2008/09 reversed the trend of large rises in the number of platelet imports.
TRALI – Platelet Apheresis
 In BC, LVPs allowed CBS to continue to meet
platelet demand with fewer imports

 The number of platelet apheresis units

collected since implementation of the TRALI
reduction measures has not decreased, in
fact increased with 38% growth
 BC Yukon data
 National growth: 6%
 National data: LVP target 34.5% actual YTD
23.4%
Summary Primary Prevention
 Reports of TRALI have decreased
since shift to male plasma
 Consistent with experience elsewhere
 Donor loss due to deferral of
previously pregnant females largely
mitigated by collection of double
platelets
Will there be enough snow
for the Olympics??

Questions?