Evidence-Based Clinical Practice Guidelines American College of Chest Physicians Heparin ...

43 Pages·2008·0.56 MB·English

Treatment and Prevention of
Heparin-Induced Thrombocytopenia:
American College of Chest Physicians
Evidence-Based Clinical Practice Guidelines
(8th Edition)

Theodore E. Warkentin, Andreas Greinacher, Andreas Koster and A.
Michael Lincoff

Chest 2008;133;340-380
DOI 10.1378/chest.08-0677
The online version of this article, along with updated information
and services can be found online on the World Wide Web at:
http://chestjournal.org/cgi/content/abstract/133/6_suppl/340S
CHEST is the official journal of the American College of Chest
Physicians. It has been published monthly since 1935. Copyright 2007
by the American College of Chest Physicians, 3300 Dundee Road,
Northbrook IL 60062. All rights reserved. No part of this article or PDF
may be reproduced or distributed without the prior written permission
of the copyright holder
(http://www.chestjournal.org/misc/reprints.shtml). ISSN: 0012-3692.
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians
Supplement
ANTITHROMBOTICANDTHROMBOLYTICTHERAPY8THED:ACCPGUIDELINES
Treatment and Prevention of
Heparin-Induced Thrombocytopenia*
American College of Chest Physicians
Evidence-Based Clinical Practice Guidelines
(8th Edition)
Theodore E. Warkentin, MD; Andreas Greinacher, MD; Andreas Koster, MD;
and A. Michael Lincoff, MD
Thischapterabouttherecognition,treatment,andpreventionofheparin-inducedthrombocytopenia
(HIT)ispartoftheAntithromboticandThrombolyticTherapy:AmericanCollegeofChestPhysicians
Evidence-BasedClinicalPracticeGuidelines(8thEdition).Grade1recommendationsarestrongand
indicate that the benefits do, or do not, outweigh risks, burden, and costs. Grade 2 suggests that
individual patient values may lead to different choices. Among the key recommendations in this
chapterarethefollowing:Forpatientsreceivingheparininwhomtheclinicianconsiderstheriskof
HIT to be >1.0%, we recommend platelet count monitoring over no platelet count monitoring
(Grade1C).Forpatientswhoarereceivingheparinorhavereceivedheparinwithintheprevious2
weeks,werecommendinvestigatingforadiagnosisofHITiftheplateletcountfallsby>50%,and/or
athromboticeventoccurs,betweendays5and14(inclusive)followinginitiationofheparin,evenif
the patient is no longer receiving heparin therapy when thrombosis or thrombocytopenia has
occurred (Grade 1C). For patients with strongly suspected (or confirmed) HIT, whether or not
complicatedbythrombosis,werecommenduseofanalternative,nonheparinanticoagulant(danap-
aroid [Grade 1B], lepirudin [Grade 1C], argatroban [Grade 1C], fondaparinux [Grade 2C], or
bivalirudin [Grade 2C]) over the further use of unfractionated heparin (UFH) or low-molecular-
weight heparin (LMWH) therapy or initiation/continuation of vitamin K antagonists (VKAs) [Grade
1B]. The guidelines include specific recommendations for nonheparin anticoagulant dosing that
differ from the package inserts. For patients with strongly suspected or confirmed HIT, we
recommendagainsttheuseofvitaminKantagonist(VKA)[coumarin]therapyuntilaftertheplatelet
count has substantially recovered (usually, to at least 150(cid:1)109/L) over starting VKA therapy at a
lower platelet count (Grade 1B); that VKA therapy be started only with low maintenance doses
(maximum,5mgofwarfarinor6mgofphenprocoumon)overhigherinitialdoses(Grade1B);and
that the nonheparin anticoagulant (eg, lepirudin, argatroban, danaparoid) be continued until the
plateletcounthasreachedastableplateau,theinternationalnormalizedratio(INR)hasreachedthe
intended target range, and after a minimum overlap of at least 5 days between nonheparin
anticoagulation and VKA therapy rather than a shorter overlap (Grade 1B). For patients receiving
VKAsatthetimeofdiagnosisofHIT,werecommenduseofvitaminK(10mgpoor5to10mgIV)
[Grade1C]. (CHEST2008;133:340S–380S)
Keywords:argatroban;bivalirudin;coumarin-inducednecrosis;danaparoid;fondaparinux;heparin-inducedthrombocyto-
penia;IgG;lepirudin;low-molecular-weightheparin;unfractionatedheparin
Abbreviations: ACT(cid:1)activated clotting time; APTT(cid:1)activated partial thromboplastin time; CI(cid:1)confidence interval;
CPB(cid:1)cardiopulmonary bypass; DIC(cid:1)disseminated intravascular coagulation; DTI(cid:1)direct thrombin inhibitor;
DVT(cid:1)deepvenousthrombosis;ECT(cid:1)ecarinclottingtime;EIA(cid:1)enzymeimmunoassay;EMEA(cid:1)EuropeanAgencyfor
theEvaluationofMedicinalProducts;FDA(cid:1)FoodandDrugAdministration;HIPA(cid:1)heparin-inducedplateletactivation
assay; HIT(cid:1)heparin-induced thrombocytopenia; INR(cid:1)international normalized ratio; LMWH(cid:1)low-molecular-weight
heparin; OD(cid:1)optical density; OR(cid:1)odds ratio; PCI(cid:1)percutaneous coronary intervention; PE(cid:1)pulmonary embolism;
PF4(cid:1)platelet factor 4; RCT(cid:1)randomized controlled trial; RR(cid:1)relative risk; SC(cid:1)subcutaneous; SRA(cid:1)serotonin
releaseassay;UFH(cid:1)unfractionatedheparin;VKA(cid:1)vitaminKantagonist
340S AntithromboticandThrombolyticTherapy8thEd:ACCPGuidelines
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians
toring at least every 2 or 3 days from day 4 to
Summary of Recommendations
day 14 (or until heparin is stopped, whichever
1.0 Recognition of HIT occurs first) over less frequent platelet count
monitoring (Grade 2C).
1.1 Platelet Count Monitoring for HIT
1.1.4 Platelet Count Monitoring in Postoperative
1.1.Forpatientsreceivingheparininwhomthe Patients Receiving UFH Antithrombotic
clinicianconsiderstheriskofHITtobe>1.0%, Prophylaxis (Highest Risk Group for HIT)
we recommend platelet count monitoring over
no platelet count monitoring (Grade 1C). For 1.1.4. For patients who are receiving postoper-
patients receiving heparin who have an esti- ative antithrombotic prophylaxis with UFH, ie,
mated risk of HIT of 0.1 to 1.0%, we suggest the patient population at highest risk for HIT
platelet count monitoring over no platelet (HIT risk >1%), we suggest at least every-
count monitoring (Grade 2C). other-day platelet count monitoring between
postoperative days 4 to 14 (or until UFH is
1.1.1 Platelet Count Monitoring of Patients stopped, whichever occurs first) over less fre-
Recently Treated With Heparin quent platelet count monitoring (Grade 2C).
1.1.5 Platelet Count Monitoring in Patients in
1.1.1. For patients who are starting UFH or
Whom HIT is Infrequent (0.1 to 1%)
LMWH treatment and who have received UFH
within the past 100 days, or those patients in
1.1.5. For medical/obstetrical patients who are
whom exposure history is uncertain, we recom-
receiving prophylactic-dose UFH, postoperative
mend obtaining a baseline platelet count and
patientsreceivingprophylactic-doseLMWH,post-
then a repeat platelet count within 24 h of
operative patients receiving intravascular cath-
starting heparin over not obtaining a repeat
eter UFH “flushes,” or medical/obstetrical pa-
platelet count (Grade 1C).
tients receiving LMWH after first receiving
UFH (estimated HIT risk, 0.1 to 1%), we sug-
1.1.2 Anaphylactoid Reactions After IV UFH Bolus
gest platelet count monitoring at least every 2
1.1.2.Forpatientsinwhomacuteinflammatory, or3daysfromday4today14(oruntilheparin
cardiorespiratory, neurologic, or other unusual is stopped, whichever occurs first), when prac-
symptoms and signs develop within 30 min tical, over less frequent platelet count monitor-
following an IV UFH bolus, we recommend ing (Grade 2C).
performing an immediate platelet count mea-
1.1.6 Platelet Count Monitoring When HIT is
surement, and comparing this value to recent
(cid:2)
Rare ( 0.1%): UFH and LMWH
prior platelet counts, over not performing a
platelet count (Grade 1C).
1.1.6. For medical/obstetrical patients who are
1.1.3 Platelet Count Monitoring in Patients receivingonlyLMWH,ormedicalpatientswho
Receiving Therapeutic-Dose UFH are receiving only intravascular catheter UFH
flushes(HITrisk<0.1%),wesuggestclinicians
do not use routine platelet count monitoring
1.1.3. For patients who are receiving therapeu-
(Grade 2C).
tic-dose UFH, we suggest platelet count moni-
1.1.7 Platelet Count Monitoring When HIT is Rare
*From McMaster University (Dr. Warkentin), Hamilton, ON, (cid:2)
( 0.1%): Fondaparinux
Canada; Institute for Immunology and Transfusion Medicine
(Dr. Greinacher), Ernst-Moritz-Arndt University Greifswald,
Greifswald, Germany; Deutsches Herzzentrum Berlin (Dr. Ko- 1.1.7. For patients who are receiving fondapa-
ster),Berlin,Germany;andClevelandClinicLernerCollegeof
rinux thromboprophylaxis or treatment, we rec-
MedicineofCaseWesternReserveUniversity(Dr.Lincoff),The
ClevelandClinicFoundation,Cleveland,OH. ommend that clinicians do not use routine plate-
ManuscriptacceptedDecember20,2007. let count monitoring (Grade 1C).
Reproductionofthisarticleisprohibitedwithoutwrittenpermission
fromtheAmericanCollegeofChestPhysicians(www.chestjournal.
1.1.8 Management of Patients in Whom Platelet
org/misc/reprints.shtml).
Correspondence to: Theodore E. Warkentin, MD, Room I-180A, Counts Are Not Monitored
HamiltonRegionalLaboratoryMedicineProgram,HamiltonHealth
Sciences,GeneralSite,237BartonStE,Hamilton,ON,L8L2X2
1.1.8. In outpatients who will receive heparin
Canada;e-mail:[email protected]
DOI:10.1378/chest.08-0677 prophylaxis or treatment, informed consent
www.chestjournal.org CHEST/133/6/JUNE,2008SUPPLEMENT 341S
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians
shouldincludeHITanditstypicalsequelae(new <90 (cid:3)mol/L), with lower infusion rates for
thrombosis, skin lesions) and the patient should patients with higher serum creatinine levels
(cid:3)
be advised to seek medical advice if these events (creatinine, 90 to 140 mol/L: starting infusion
occur (Grade 2C). rate, 0.05 mg/kg/h; creatinine, 140 to 400
(cid:3)
mol/L: starting infusion rate, 0.01 mg/kg/h;
1.1.9 Screening for Subclinical HIT Antibody creatinine>400(cid:3)mol/L:startinginfusionrate,
Seroconversion 0.005 mg/kg/h) [Grade 1C]. Furthermore, we
recommend that the initial IV bolus either be
1.1.9. In patients who receive heparin, or in omitted or, in case of perceived life- or limb-
whom heparin treatment is planned (eg, for threatening thrombosis, be given at a reduced
cardiac or vascular surgery), we recommend dose (0.2 mg/kg) [Grade 1C]. Further, we rec-
against routine HIT antibody testing in the ommend that APTT monitoring be performed
absenceofthrombocytopenia,thrombosis,hep- at 4-h intervals until it is apparent that steady
arin-induced skin lesions, or other signs point- state within the therapeutic range (1.5- to 2.0-
ing to a potential diagnosis of HIT (Grade 1C). times patient baseline [or mean laboratory]
APTT) is achieved (Grade 1C).
2.1.3.Whenargatrobanisusedtotreatpatients
1.1.10 When Should HIT Be Suspected?
who have heart failure, multiple organ system
failure, or severe anasarca or who are postcar-
1.1.10. For patients who are receiving heparin
diac surgery, we suggest beginning the initial
or have received heparin within the previous 2
(cid:3)
infusion at a rate between 0.5 and 1.2 g/kg/
weeks, we recommend investigating for a diag-
> min, with subsequent adjustments using the
nosisofHITiftheplateletcountfallsby 50%,
APTT, over the usual recommended starting
and/orathromboticeventoccurs,betweendays
(cid:3)
dose of 2.0 g/kg/min (Grade 2C).
5 and 14 (inclusive) following initiation of hep-
2.1.4.Whendanaparoidisusedtotreatpatients
arin, even if the patient is no longer receiving
withstronglysuspected(orconfirmed)HIT,we
heparin therapy when thrombosis or thrombo-
recommend a therapeutic-dose regimen (see
cytopenia has occurred (Grade 1C).
text) administered (at least initially) by the IV
routeoverprophylactic-doseregimensorinitial
1.2 Special Situation: Anticoagulant Prophylaxis
SC administration (Grade 1B).
and Platelet Count Monitoring After Cardiac
2.1.5. For patients with strongly suspected or
Surgery
confirmed HIT, whether or not there is clinical
evidence of lower-limb DVT, we recommend
1.2. For postoperative cardiac surgery patients,
routineultrasonographyofthelower-limbveins
werecommendinvestigatingforHITantibodiesif
> for investigation of DVT over not performing
the platelet count falls by 50%, and/or a throm-
routine ultrasonography (Grade 1C).
botic event occurs, between postoperative days 5
(cid:2)
and 14 (inclusive; day of cardiac surgery day 0) 2.2 VKAs
[Grade1C].
2.2.1 Management of Direct Thrombin
Inhibitor–VKA Overlap
2.0 Treatment of HIT
2.1 Nonheparin Anticoagulants for Treating HIT 2.2.1. For patients with strongly suspected or
(With or Without Thrombosis) confirmedHIT,werecommendagainsttheuseof
VKA (coumarin) therapy until after the platelet
2.1.1. For patients with strongly suspected (or count has substantially recovered (ie, usually to at
confirmed) HIT, whether or not complicated by least150(cid:1)109/L)overstartingVKAtherapyata
thrombosis,werecommenduseofanalternative, lower platelet count (Grade 1B); that VKA ther-
nonheparinanticoagulant(danaparoid[Grade 1B], apybestartedonlywithlow,maintenancedoses
lepirudin [Grade 1C], argatroban [Grade 1C], (maximum, 5 mg of warfarin or 6 mg of phen-
fondaparinux [Grade 2C], bivalirudin [Grade procoumon) rather than with higher initial
2C]) over the further use of UFH or LMWH doses (Grade 1B); and that the nonheparin anti-
therapy or initiation/continuation of a VKA coagulant (eg, lepirudin, argatroban, danap-
(Grade 1B). aroid) be continued until the platelet count has
2.1.2. For patients receiving lepirudin, the ini- reached a stable plateau, the INR has reached
tial lepirudin infusion rate should be no higher theintendedtargetrange,andafteraminimum
than 0.10 mg/kg/h (patients with creatinine overlap of at least 5 days between nonheparin
342S AntithromboticandThrombolyticTherapy8thEd:ACCPGuidelines
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians
anticoagulation and VKA therapy rather than a ogy have been adapted to the unique features of
shorter overlap (Grade 1B). bivalirudin pharmacology) [Grade 1B] or during
“off-pump”cardiacsurgery(Grade1B);usinglepi-
2.2.2 Reversal of VKA Anticoagulation rudinforintraoperativeanticoagulation(ifECTis
available and patient has normal renal function
2.2.2. For patients receiving a VKA at the time and is judged to be at low risk for postcardiac
of diagnosis of HIT, we recommend use of surgeryrenaldysfunction)[Grade2C];usingUFH
vitamin K (10 mg po or 5 to 10 mg IV) [Grade plus the antiplatelet agent epoprostenol (if ECT
1C]. monitoring is not available or renal insufficiency
precludes lepirudin use) [Grade 2C]; using UFH
2.3 LMWH for HIT plus the antiplatelet agent, tirofiban (Grade 2C);
orusingdanaparoid for intraoperative antico-
2.3.1.ForpatientswithstronglysuspectedHIT, agulation for off-pump coronary artery by-
whether or not complicated by thrombosis, we pass surgery (Grade 2C) over performing the
recommend against use of LMWH (Grade 1B). surgery with UFH when platelet-activating
anti-PF4/heparin antibodies are known to be
present in a patient with acute or recent HIT.
2.4 Prophylactic Platelet Transfusions for HIT
3.2.2. For patients with subacute HIT (plate-
let count recovery, but continuing HIT anti-
2.4.1. For patients with strongly suspected or
body positive), we recommend delaying sur-
confirmed HIT who do not have active bleed-
gery (if possible) until HIT antibodies
ing,wesuggestthatprophylacticplatelettrans-
(washed platelet activation assay) are nega-
fusions should not be given (Grade 2C).
tive, then using heparin (see Recommenda-
tion 3.1.1.) over using a nonheparin anticoag-
3.0 Special Patient Populations
ulant (Grade 1C). If surgery cannot be
3.1 Patients With Previous HIT Undergoing delayed, we suggest the use of a nonheparin
Cardiac or Vascular Surgery anticoagulant (see Recommendation 3.2.1.)
over the use of UFH (Grade 2C).
3.1.1.ForpatientswithahistoryofHITwhoare
HITantibodynegativeandrequirecardiacsur-
3.3 Percutaneous Coronary Interventions
gery, we recommend the use of UFH over a
nonheparin anticoagulant (Grade 1B).
3.3.1. For patients with strongly suspected (or
3.1.2. For patients with a history of HIT who are
confirmed) acute HIT who require cardiac
antibody positive by platelet factor 4 (PF4)-depen-
catheterizationorpercutaneouscoronaryinter-
dentenzymeimmunoassay(EIA)butantibodyneg-
vention (PCI), we recommend a nonheparin
ativebywashedplateletactivationassay,werecom-
anticoagulant (bivalirudin [Grade 1B], argatro-
mend the use of UFH over a nonheparin
ban [Grade 1C], lepirudin [Grade 1C], or dan-
anticoagulant(Grade2C).
aparoid [Grade 1C]) over UFH or LMWH
Remark:Preoperativeandpostoperativeanticoag-
(Grade 1B).
ulation, if indicated, should be given with a nonhe-
3.3.2. For patients with previous HIT (who
parin anticoagulant.
are antibody negative) who require cardiac
catheterization or PCI, we suggest use of a
3.2 Patients With Acute or Subacute HIT
nonheparin anticoagulant (see Recommenda-
Undergoing Cardiac Surgery
tion 3.3.1.) over UFH or LMWH (Grade 2C).
3.2.1.ForpatientswithacuteHIT(thrombocyto-
penic,HITantibodypositive)whorequirecardiac Heparin-induced thrombocytopenia (HIT) is an
surgery, we recommend one of the following antibody-mediated adverse effect of heparin
alternativeanticoagulantapproaches(indescend- that is important because of its strong association
ing order of preference): delaying surgery (if with venous and arterial thrombosis.1–4 Patients
possible) until HIT has resolved and antibodies treated with heparin in whom HIT develops consti-
arenegative(thenseeRecommendation3.1.1.)or tuteacohortwithsubstantiallyincreasedthrombotic
weaklypositive(thenseeRecommendation3.1.2.) risk,bothinrelative(oddsratio[OR]forthrombosis,
[Grade 1B]; using bivalirudin for intraoperative 20 to 401–5) and absolute (thrombosis risk, 30 to
anticoagulation during cardiopulmonary bypass 75%1–10)terms,dependingonthepatientpopulation
(iftechniquesofcardiacsurgeryandanesthesiol- affected.
www.chestjournal.org CHEST/133/6/JUNE,2008SUPPLEMENT 343S
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians
Because the diagnosis is based on both clinical and nitude of a positive EIA test result, the greater the
serologic grounds, clinicians should consider HIT a likelihood that the patient has HIT, given a certain
clinicopathologic syndrome.11–14 Thus, neither throm- pretest probability. However, a very strong EIA test
bocytopenia or thrombosis without the presence of result does not necessarily mean that platelet-
heparin-dependent antibodies, nor the isolated pres- activating IgG antibodies are present; conversely,
ence of antibodies without thrombocytopenia, throm- only about 5 to 10% of sera showing reactivity 0.4
bosis, or other clinical sequelae, meet the criteria for to 1.0 OD U in an EIA nonetheless contain strong
HIT.Rather,cliniciansmakeadiagnosisofHITwhen platelet-activating antibodies.29
any of the following events occurs in association with This chapter is organized into sections on the
the presence of “HIT antibodies” detected by in vitro recognition, treatment, and prevention of HIT. The
assays:(1)anotherwiseunexplainedplateletcountfall scope of our recommendations include platelet
(definedbyvariousinvestigatorsasaminimumplatelet count monitoring for HIT as well as management of
count fall of 30%,15,16 40%,17 or 50%2—even if the HIT,bothinpatientsdetectedbythrombocytopenia
plateletcountnadirremains(cid:3)150(cid:4)109/L(note:the alone (“isolated HIT”) and in patients who present
“baseline” platelet count is not necessarily the prehe- with HIT-associated thrombosis. The interrelated-
parin platelet count, but rather the highest platelet ness of platelet count monitoring and treatment
count during the 2-week period that follows initiation recommendations is clear when one considers that
of heparin therapy and that immediately precedes the “isolatedHIT”(apatientpopulationwithsubstantial
platelet count decline indicating HIT); (2) venous or riskofthrombosis)bydefinitioncanbedetectedonly
arterialthrombosis(mostoften,deepvenousthrombo- by platelet count monitoring. Furthermore, even in
sis [DVT], pulmonary embolism [PE], limb artery patientswiththrombosiscomplicatingHIT,theavailabil-
thrombosis, thrombotic stroke, myocardial infarction, ity of serial platelet counts often provides the key
adrenal hemorrhagic necrosis [indicating adrenal vein informationtopromptconsiderationofthediagnosis
thrombosis]); (3) skin lesions at heparin injection of HIT. Table 1 lists the inclusion and exclusion
sites18; or (4) acute systemic (anaphylactoid) reactions criteriaforthestudiesusedtoformulateourrecom-
(eg, fever/chills, tachycardia, hypertension, dyspnea, mendations.
cardiopulmonary arrest) that occur after IV heparin
bolus administration.7 Diagnostic specificity can be
further increased by use of a sensitive washed platelet 1.0 Recognition of HIT
activation assay; a positive platelet activation assay is
1.1 Platelet Count Monitoring for HIT
much more specific for clinical HIT than a positive
platelet factor 4 (PF4)-dependent immunoassay.19–23 HIT occurs most commonly in certain patient
Withsuchaclinicopathologicviewpoint,clinicianscan populations, such as postoperative patients who re-
diagnose clinical HIT even when the patient’s platelet ceive standard, unfractionated heparin (UFH) for
(cid:1)
count does not fall by as much as 30% (eg, a patient 1week(forreview,seeLeeandWarkentin6).One
with heparin-induced necrotizing skin lesions and ad- definitionclassifiesanadversereactionas“common”
(cid:3)
renal necrosis associated with formation of platelet- if its incidence is 1%.30 In other clinical settings,
activating anti-PF4/heparin antibodies18).21,24 Indeed, the estimated risk of HIT can be described as
(cid:2)
in about 25% of HIT patients, a thrombotic event “uncommon” (0.1 to 1%) or “rare” ( 0.1%).30 As
during heparin treatment precedes the subsequent describedlater,thereisevidencethatinitialisolated
HIT-associated platelet count fall.1,16 HIThasasubstantialriskofevolvingtosymptomatic
The neoepitopes recognized by HIT antibodies and fatal thrombosis. Further, prospective cohort
arelocatedonPF4,andareformedwhenPF4binds studies (with historical controls) suggest that anti-
to heparin.25–27 HIT antibodies activate platelets thrombotictherapyreducestheriskofthrombosisin
intravascularly, causing release of platelet micropar- patientswithisolatedHIT.Inaddition,HITcanlead
ticles and increased thrombin generation. However, to life- and limb-threatening complications, a risk
onlyasubsetofanti-PF4/heparinantibodiesactivate that could increase with delay in diagnosis or increase
platelets,19,20,28whichexplainsthegreaterdiagnostic inheparindose(totreatunrecognizedHIT-associated
specificity of certain platelet activation assays (eg, thrombosis),orthroughuseofwarfarin.Theseconsid-
platelet serotonin release assay [SRA], heparin- erationssuggestthatroutineplateletcountmonitoring
induced platelet activation (HIPA) assay] for HIT for HIT is appropriate in at least some clinical situa-
compared with the PF4-dependent EIA.19,20,23,29 tions,andthatthegreatertheriskofHIT,thestronger
There is a correlation between the degree of reac- the rationale for regular monitoring.
tivity in the EIA, expressed in optical density (OD) Another consideration that supports a role for
units, and the presence of platelet-activating anti- platelet count monitoring is that HIT antibody sero-
PF4/heparin antibodies. Thus, the greater the mag- conversion and resulting “typical-onset” HIT usually
344S AntithromboticandThrombolyticTherapy8thEd:ACCPGuidelines
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians
Table1—QuestionDefinitionandStudyEligibilityCriteria(Section:Introduction)
InclusionCriteria
Exclusion
Section Population Intervention(s)orExposure Outcome Methodology Criteria
1.1 Hospitalizedmedical, UFHorLMWHfor5–20d HIT(plateletcountdecrease Metaanalyses,RCTs, (cid:2)25patients
surgical,or vsnoanticoagulation (cid:3)30%)and/ornew cohortstudies
obstetricpatients thrombosis/skinlesions
(secondarytoplatelet-
activating
anti-PF4/heparin
antibodies);
anti-PF4/heparinantibody
formation(platelet
activationassaysand/or
PF4/polyanionEIA)
1.2 Cardiacsurgery Noanticoagulationvs HIT;HIT-associatedVTE; Metaanalyses;RCTs; (cid:2)50patients
patients anticoagulation(UFH, HITantibodyformation cohortstudies
undergoing LMWH,fondaparinux)
on-pumpand aftersurgery
off-pumpsurgery
2.1 HITpatients(with Non-heparinanticoagulation Newthrombosis;mortality; RCTs;cohortstudies (cid:2)25patients(no
orwithout (lepirudin,argatroban, limbamputation; exclusionsfor
thrombosis) bivalirudin,danaparoid, compositeofabove;drug reportson
orfondaparinux) anaphylaxis drug-associated
anaphylaxis)
2.2 HITpatients(with VKAs(coumarins) VKAtherapy-associated Cohortstudies Patientswithout
orwithout thrombosis,including HITantibodies
thrombosis) venouslimbgangrene,
skinnecrosis
2.3 HITpatients(with LMWH Thrombosis;plateletcount Cohortstudies (cid:2)25patients
orwithout recovery
thrombosis)
2.4 HITpatients(with Platelettransfusionsvsno Newthrombosis Cohortstudiesandcase (cid:2)5patients
orwithout platelettransfusions series
thrombosis)
3.1 Patientswith Repeatheparinexposure RecurrenceofHIT;repeat Prospectivecohortstudies; (cid:2)5patients
previousHIT formationofHIT retrospectivecohort
undergoingcardiac antibodies studies
surgery
3.2 Patientswithacute Anticoagulantapproaches Proceduralsuccess(as Prospectivecohortstudies; (cid:2)5patients
orsubacuteHIT duringcardiacsurgery: definedbystudyauthors) retrospectivecohort
undergoingcardiac lepirudin,argatroban, studies
surgery danaparoid,bivalirudin,
UFHplusepoprostenol,
UFHplustirofiban
3.3 Patientswith Non-heparinanticoagulation Proceduralsuccess(as Prospectivecohortstudies; (cid:2)5patients
previousoracute definedbystudyauthors) retrospectivecohort
HITundergoing studies
PCI
3.4 Patientswithacute Non-heparinanticoagulation Proceduralsuccess(as Retrospectivecohort (cid:2)5patients
HITundergoing definedbystudyauthors) studies
hemodialysis
4.1 Hospitalizedmedical, ComparisonofUFHand FrequencyofHIT(using Metaanalyses;RCTs; (cid:2)100patients
surgical,or LMWH/fondaparinux explicitcriteria);frequency nonrandomized
obstetricpatients; ofHITantibodyformation controlledstudies;
cardiacsurgery retrospectivecohort
patients studies
occur during specific time periods following initia- when the first day of the immunizing heparin expo-
tion of heparin, namely days 5 to 10 (seroconversion sureisconsideredtobe“day0.”1,2,6,7,31,32(Day4can
and initial platelet count fall) and days 7 to 14 be included within the period of platelet count
(reaching a threshold defining thrombocytopenia), monitoring for HIT because it can provide a com-
www.chestjournal.org CHEST/133/6/JUNE,2008SUPPLEMENT 345S
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians
parativepreimmunizationreferencepoint.)Further, toring over no platelet count monitoring
“rapid-onset HIT” (in which the platelet count fall (Grade 1C). For patients receiving heparin
begins within 24 h of starting heparin) is strongly who have an estimated risk of HIT of 0.1 to
associated with recent heparin exposure (within the 1.0%, we suggest platelet count monitoring
past 100 days, and especially the last 30 days).31,32 over no platelet count monitoring (Grade 2C).
The frequency of HIT among patients exposed to
heparin is highly variable, and is influenced by the 1.1.1 Platelet Count Monitoring of Patients
(cid:3) (cid:3)
heparin preparation (bovine UFH porcine UFH Recently Treated With Heparin
low-molecular-weightheparin[LMWH]),19,20,33–41the
“Rapid-onset HIT” refers to patients who have a
type of heparin-exposed patient population (postsur-
(cid:3) (cid:3) large platelet count fall attributable to HIT antibod-
gery medical pregnancy),19,20,41–46 duration of
(cid:3) ies within 24 h of starting heparin.31,32 Contrary to
heparin exposure,7 and patient sex (female male).41
popular assumption, this phenomenon is not caused
Thus, whether to perform platelet count monitoring,
by an anamnestic immune response, but rather
andtheintensityofsuchmonitoring,dependsonthese
results from the administration of heparin to a
considerations, particularly heparin and patient type
patient who has already-circulating HIT antibodies
and the duration of heparin use. Therefore, it is
thatresultedfromarecentheparinexposure.31,32As
appropriate to perform platelet count monitoring in
a general rule, exposure within the past 100 days
certain clinical situations, and to focus platelet count
(and especially within the last month) is associated
monitoring during those times when HIT usually oc-
with the phenomenon of rapid-onset HIT.
curs. Furthermore, the rationale for platelet count
monitoring is stronger when monitoring is relatively
easy (ie, in a hospital inpatient), and weaker when Recommendation
monitoring is more difficult (ie, out-patient settings.
Table 2 summarizes the various risk factors for HIT, 1.1.1. For patients who are starting UFH or
and classifies risk into three different groups: high LMWH treatment and who have received
(cid:3)
( 1.0%), intermediate (0.1 to 1.0%), and low UFH within the past 100 days, or those pa-
(cid:2)
( 0.1%).1,2,4,6,15,17,19,20,34–71 tients in whom exposure history is uncertain,
A potential downside of platelet count monitor- we recommend obtaining a baseline platelet
ing is that patients with a decrease in platelet count and then a repeat platelet count within
countsforreasonsotherthanHITmaybewrongly 24 h of starting heparin over not obtaining a
suspected of having this diagnosis. As the alterna- repeat platelet count (Grade 1C).
tive nonheparin anticoagulants have a relatively
high risk of bleeding, there is the potential for
1.1.2 Anaphylactoid Reactions After IV UFH Bolus
treatment-relatedadverseeventsasaconsequence
of platelet count monitoring. Rarely, patients develop acute inflammatory (eg,
fever, chills) or cardiorespiratory (eg, hypertension,
UnderlyingValuesandPreferences:Thefollowing tachycardia, dyspnea, chest pain, cardiorespiratory
recommendations regarding monitoring of platelet arrest) symptoms and signs within 30 min following
count share the same underlying values and prefer- an IV heparin bolus.7,47 Also termed acute systemic
ences,asfollows.Therecommendationsplaceahigh reactions, these can also mimic acute PE (pseudo-
value on the possible benefits of early diagnosis and pulmonary embolism48) and strongly suggest acute
consequent early treatment of HIT to prevent se- in vivo platelet activation secondary to HIT. This
quelae and a lower value on the burden and cost of presentation mandates a prompt platelet count
monitoring platelet counts, including the conse- measurement, as an abrupt platelet count fall in
quencesoffurtherinvestigationandmanagementof this clinical context supports the diagnosis of HIT.
the high proportion of patients with a significant fall The platelet count drop is frequently transient,2
in platelet count who do not have HIT and the risks andthusadelayindeterminingtheplateletcount,
associated with unnecessary withdrawal of heparin especially if heparin is stopped, may result in a
and unnecessary use of alternative agents with a missed diagnosis.
higher bleeding risk.
Recommendation
Recommendation
1.1.2.Forpatientsinwhomacuteinflammatory,
1.1. For patients receiving heparin in whom cardiorespiratory, neurologic, or other unusual
the clinician considers the risk of HIT to be symptoms and signs develop within 30 min
> 1.0%, we recommend platelet count moni- following an IV UFH bolus, we recommend
346S AntithromboticandThrombolyticTherapy8thEd:ACCPGuidelines
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians
Table2—RiskFactorsforHIT:ImplicationsforPlateletCountMonitoring(Section:1.1)*
RelativeImportanceofRiskFactor
RiskFactors Major(OR(cid:3)5) Moderate(OR3–5) Minor(OR1–3)
Heparinduration(cid:3)4d† Yes
Recentheparin(past100d)‡ Yes
UFH(cid:3)LMWH§ Yes
Postsurgery(cid:3)medical(cid:3)obstetric Yes
Doseofheparin
Immunizing:prophylaxis(cid:3)therapeutic(cid:1) Yes
Manifesting:therapeutic(cid:3)prophylaxis Yes
(cid:3)“flushes”¶
Gender:female(cid:3)male Yes
Examplesofpatientgroupswithriskestimatedtobe(cid:3)1%
Postoperativepatientsreceivingprophylactic-doseUFH(cid:3)4d
Postoperativepatientsreceivingtherapeutic-doseUFH(cid:3)4d#
Examplesofpatientgroupswithriskestimatedtobe0.1–1%
Medical/obstetricpatientsreceivingprophylacticortherapeutic-doseUFH(cid:3)4d
PostsurgerypatientsreceivingLMWH(cid:3)4d
PostsurgerypatientsreceivingUFH“flushes”(cid:3)4d
Medical/obstetricpatientsreceivingLMWHafterfirstreceivingUFH
Examplesofpatientsgroupswithriskestimatedtobe(cid:2)0.1%
Medical/obstetricpatientsreceivingLMWH(cid:3)4d**
Medical/obstetricpatientsreceivingonlyheparinflushes
AnypatientreceivingUFHorLMWH(cid:2)4d
*Baseduponassessmentofthepublishedliterature.1,2,4,6,15,17,19,20,34–71
†Riskdeclinesafter14days(intheabsenceofinterveningsurgery).
‡Riskofrapid-onsetHITifheparinisrestartedinapatientexposedwithinthepast100days(andespeciallythelast30days).
§DifferenceinriskbetweenUFHandLMWHbestestablishedinpostsurgerypatientsandinfemales.
(cid:1)Theoretically,stoichiometricconcentrationsofPF4/UFHandPF4/LMWHaremostlikelytobeachievedatprophylacticdoses.
¶AmongpatientswhohaveHITantibodies,higherdosesofheparinusuallyresultingreaterplateletcountfalls.
#Bestestablishedinpost-cardiacsurgerypatients.
**Onestudy38suggestedthatthefrequencyofHITinmedicalpatientsreceivingLMWHcouldbebetween0.1–1%,butthisstudyisastatistical
outlieranditsconclusionsremaintobeconfirmed.39,41
performing an immediate platelet count mea- toring at least every 2 or 3 days from day 4 to
surement, and comparing this value to recent day 14 (or until heparin is stopped, whichever
prior platelet counts, over not performing a occurs first) over less frequent platelet count
platelet count (Grade 1C). monitoring (Grade 2C).
1.1.4 Platelet Count Monitoring in Postoperative
1.1.3 Platelet Count Monitoring in Patients
Patients Receiving UFH Antithrombotic
Receiving Therapeutic-Dose UFH
Prophylaxis (Highest Risk Group for HIT)
ForpatientsreceivingporcineUFHintherapeutic
PatientgroupsatthehighestriskofHIT(1to5%)
doses, by either the IV or subcutaneous (SC) route,
include postoperative orthopedic, cardiac, and vas-
for the treatment of venous or arterial thrombosis,
cular surgery patients who are receiving UFH for 1
the risk of HIT has been estimated to be at most
to2weeks,15,17,19,20,33,63–67and,likely,otherpostsur-
about 1%,6 based on a review of studies of the
gery patient populations.
frequencyofHITinpatientsreceivingporcineUFH
forvenousthromboembolism.4,36,49–62However,the Recommendation
two most recent studies4,36 identified only 1 patient
with HIT among 594 treated with therapeutic-dose 1.1.4. For patients who are receiving postoper-
UFH, suggesting that the frequency among this ative antithrombotic prophylaxis with UFH, ie,
(cid:2)
patient population is likely 1%. the patient population at highest risk for HIT
(HIT risk >1%), we suggest at least every-
Recommendation other-day platelet count monitoring between
postoperative days 4 to 14 (or until UFH is
1.1.3. For patients who are receiving therapeu- stopped, whichever occurs first) over less fre-
tic-dose UFH, we suggest platelet count moni- quent platelet count monitoring (Grade 2C).
www.chestjournal.org CHEST/133/6/JUNE,2008SUPPLEMENT 347S
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians
While we gave a strong recommendation 1.1.6 Platelet Count Monitoring When HIT is Rare
(cid:2)
overall in favor of performing platelet count ( 0.1%): UFH and LMWH
monitoring when the risk of HIT was judged to
In medical and obstetrical patients receiving
>
be 1%(seeRecommendation1.1),ourrecom-
LMWH, the risk of HIT appears to be rare
mendation for the specific intensity of platelet (cid:2)
( 0.1%).36,37,39,42–45Forexample,onlyonepossible
countmonitoringinthisandotherpatientpop-
case43 of HIT was observed among 1,167 pregnan-
ulations (see also Recommendations 1.1.3 and
cies treated with LMWH in three studies42–44; a
1.1.5) have been given a weak (Grade 2) recom-
more recent review45 of LMWH use during 2,777
mendation because no study exists comparing pregnancies identified no cases of HIT. Although
outcomes using any particular platelet count fewer data exist with respect to medical patients
monitoringstrategy.Oursuggestiontoperform receiving LMWH or UFH as flushes (eg, oncology
every-other-day monitoring takes into account patients with indwelling catheters),72,73 the experi-
the observation that platelet count declines in enceoftheauthorsisthatHITisrareinthissetting.
HIT,whentheyoccur,arerelativelyrapid(me- However, one prospective, multicenter study38 of
dian of 2 to 3 days from the postoperative peak LMWH administered to medical patients reported
to a >50% platelet count decline).1,2,7 thefrequencyofHITtobe0.8%.Limitationsofthis
study suggest, however, that it provides an overesti-
mate of HIT incidence: (1) not all patients under-
1.1.5 Platelet Count Monitoring in Patients in
went testing with the more specific platelet activa-
Whom HIT is Infrequent (0.1 to 1%)
tion assays, and (2) some of the clinical features (eg,
Thereareseveralpatientgroupsinwhichtheriskof early onset of thrombocytopenia) were not charac-
HIT can be classified as “uncommon” (ie, 0.1 to 1%). teristic of this adverse drug reaction. Furthermore,
These include medical (including patients with acute thedatarepresentastatisticaloutliercomparedwith
coronary syndrome) or obstetrical patients receiving otherstudies,39,41andinourviewfurtherstudiesare
prophylactic-dose UFH4,6,46,60–62,68–70; postoperative required before recommending that platelet count
patients receiving LMWH1,2,6,17,19,20,63,66; postopera- monitoring be performed routinely in medical pa-
tive/critical care patients receiving UFH flushes6,71; tients receiving LMWH.
and, theoretically, medical patients receiving LMWH
after having received one or more preceding doses of Recommendation
UFH. In some settings (eg, patients receiving outpa-
tientLMWH),itmaybeimpracticaltoobtainplatelet 1.1.6. For medical/obstetrical patients who are
counts. Thus, less frequent (or no) platelet count receivingonlyLMWH,ormedicalpatientswho
monitoringmaybeappropriateinthesepatients,espe- are receiving only intravascular catheter UFH
ciallyiftheriskisthoughttobecloserto0.1%than1% flushes(HITrisk<0.1%),wesuggestclinicians
(eg,postoperativepatientsreceivingLMWH)andifthe do not use routine platelet count monitoring
patientisinstructedtocontactthephysicianpromptlyif (Grade 2C).
signs or symptoms of venous thromboembolism (the Underlying Values and Preferences: This recom-
most common complication of HIT) occur or painful mendationplacesalowervalueontherarediagnosis
skin lesions develop at the heparin injection sites (see and early treatment of HIT to prevent sequelae and
also Recommendation 1.1.8). ahighervalueontheburdenandcostofmonitoring
platelet counts.
Recommendation
1.1.7 Platelet Count Monitoring When HIT is Rare
(cid:2)
( 0.1%): Fondaparinux
1.1.5. For medical/obstetrical patients who
are receiving prophylactic-dose UFH, postop- Fondaparinux is an indirect (antithrombin-
erative patients receiving prophylactic-dose mediated) inhibitor of factor Xa modeled after the
LMWH, postoperative patients receiving intra- pentasaccharideregionofheparin.Asyndromeresem-
vascular catheter UFH “flushes,” or medical/ bling immune HIT was not reported in any of the
obstetrical patients receiving LMWH after first regulatorytrialsforthisanticoagulantinvariousclinical
receiving UFH (estimated HIT risk, 0.1 to 1%), settings(eg,orthopedicsurgery,74treatmentofvenous
we suggest platelet count monitoring at least thromboembolism75,76). However, anti-PF4/heparin
every 2 or 3 days from day 4 to day 14 (or until antibodies—somewithplatelet-activatingproperties—
heparin is stopped, whichever occurs first), have been observed to occur in association with
when practical, over less frequent platelet fondaparinux thromboprophylaxis in a frequency simi-
count monitoring (Grade 2C). lar to that seen with the LMWH enoxaparin.77,78
348S AntithromboticandThrombolyticTherapy8thEd:ACCPGuidelines
Downloaded from chestjournal.org on December 30, 2008
Copyright © 2008 by American College of Chest Physicians