UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)
UvA-DARE (Digital Academic Repository)
The diagnosis and prognosis of venous thromboembolism : variations on a
theme
Gibson, N.S.
Publication date
2008
Link to publication
Citation for published version (APA):
Gibson, N. S. (2008). The diagnosis and prognosis of venous thromboembolism : variations
on a theme.
General rights
It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).
Disclaimer/Complaints regulations
If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.
Clinicalusefulnessofprothrombin
fragment1+2inpatientswithsuspected
pulmonaryembolism
NADINES.GIBSON,MAAIKESÖHNE,VICTORE.A.GERDES, JOOSTC.M.MEIJERS,HARRYR.BÜLLER
SUBMITTED
96
A
BSTRACT
Background
With the use of Ddimer assays in patients with clinically suspected pulmonary embolism (PE), approximately 30% can safely be ruled out without the need for further imaging. Improvement of this percentage is desirable. Therefore, we evaluatedthediagnosticaccuracyoftheprothrombinfragment1+2(F1+2),either aloneorincombinationwithDdimer.
Methods
Baseline blood samples for F1+2 and Ddimer were obtained in consecutive patients with suspected PE in a large diagnostic management study. We determinedthediagnosticaccuracyofF1+2forseveralcutofflevelsandcompared the performance to Ddimer with a receiver operator characteristic analysis. Finally,weevaluatedwhetherF1+2couldincreasetheproportionofpatientsthat safelycanbewithheldfromadditionalimagingtestinginthesubgroupofpatients withanabnormalDdimer.
Results
F1+2 and Ddimer results were available in 373 patients with suspected PE. The prevalenceofPEwas20%.AtthepredefinedF1+2cutoffvalueof0.5nmol/Lthe sensitivity, specificity and negative predictive value (NPV) with 95% confidence intervalsofF1+2were95%(8899%),26%(2130%)and95%(8898%),respectively. By decreasing the cutoff for F1+2 the sensitivity approached 100%, and the specificity fell below 13%. The area under the curve for the F1+2 assay was 0.69 (95%CI0.630.76),andfortheDdimerassay0.82(95%CI0.770.87;p<0.05).Ofthe total study population, 64% had an abnormal Ddimer of which 8% (20 patients) hadanormalF1+2testresult.ThediagnosticaccuracyofF1+2inthissubgroupof patients was comparable to the overall group, in one patient PE was confirmed (5%;95%CI0.125).
Conclusion
In conclusion, the diagnostic accuracy of F1+2 in patients with suspected PE is inferior to the currently used Ddimer test. Furthermore, the additional value of F1+2islimitedinthesubgroupofpatientswithanabnormalDdimer.
Clinicalusefulnessofprothrombinfragment1+2
97 C HAPTER 7I
NTRODUCTION
InthelastdecadeDdimertestinghasgainedaprominentroleinthediagnosticwork up of patients with suspected pulmonary embolism. Numerous studies have shown that with the use of a Ddimer assay in combination with clinical probability assessment,approximately30%ofpatientswithsuspectedpulmonaryembolismcan be safely withheld from further imaging testing and do not need anticoagulanttherapy1,2. The major drawback of this assay is that an abnormal test result is
frequentlyobservedduetootherdiseasesthanpulmonaryembolism,whichresultsin amoderatespecificityofthetest3. Theprothrombinfragment1+2(F1+2)directlyreflectsthrombingeneration,sincethis fragmentisgeneratedduringtheconversionofprothrombintothrombin.Therefore, F1+2maybeamorespecifictestforcoagulationactivationthanDdimer,whichisa reflectionoffibrindegradation.Resultsofpreviousstudiesonthediagnosticaccuracy of F1+2 in patients with venous thromboembolism have been conflicting, varying from a better to an inferior performance of F1+2, which in part can be explained by
differencesinstudymethodologyandsamplesize48.
We, therefore, assessed in a large cohort of consecutive patients with suspected pulmonary embolism, whether F1+2 might be clinically useful to rule out PE, by determiningthesensitivity,specificityandnegativepredictivevalue(NPV)ofF1+2at variouscutoffvalues.Furthermore,wecomparedthereceiveroperatorcharacteristic (ROC)curvetothatofDdimerinthesamepatientpopulation.Finally,weanalyzed whether F1+2 may have an additional value in the subgroup of patients with an abnormalDdimerandwhetheritmightincreasetheproportionofpatientsinwhom pulmonaryembolismcanbesafelyruledout.
M
ETHODS
Patients
Data were derived from a large management study that included 631 patients with
clinicallysuspectedpulmonaryembolismthathasbeenreportedpreviously9.Forthe
presentanalysisdatafrom twohospitalswereusedthatroutinelyobtainedbloodin all patients. Patients were excluded if they had received vitamin K antagonists or heparin in a therapeutic dose for more than 24 hours, had already undergone objectivetestingforvenousthromboembolism,werepregnant,wereyoungerthan18
98
yearsofage,hadanindicationforthrombolysis,orifwritteninformedconsentcould notbeobtained.ThestudyprotocolwasapprovedbytheInstitutionalReviewBoards.
Diagnostic strategy
Theprobabilityofpulmonaryembolismwasestimateduponreferralbytheattending physician based on medical history, physical examination and if available, chest X ray,ECGandbloodgasanalysis.Thephysiciancategorizedthepatientsinoneofthe following pretestprobability categories: <20%, 2050%, 5180% or >80%. Subsequently blood was drawn for Ddimer and F1+2 measurements (Tinaquant D dimer,RocheDiagnostica,Mannheim,GermanyandEnzygnostF1+2,DadeBehring, Marburg, Germany, respectively). Pulmonary embolism was considered to be excluded if the clinical probability of pulmonary embolism was less than 20% combinedwithaDdimertestresultbelow0.5mg/L.Inallotherpatientsaventilation perfusion lung scan was obtained. These were classified as normal (no pulmonary embolism), high probability (pulmonary embolism), or nondiagnostic as defined
earlier10.Patientsinthelattercategoryunderwentbilateralserialultrasonographyof
thelegondays1,3or4and7.Anabnormalultrasoundwasconsideredtoconfirmthe presence of pulmonary embolism. Hence, pulmonary embolism was considered excludediftheclinicalprobabilitywas<20%incombinationwithanormalDdimer, ifthelungscanwasnormal,or,inthepatientswithanondiagnosticscaniftheserial ultrasoundswerenormal.Alternativelypulmonaryembolismwasdiagnosedinthose withahighprobabilitylungscintigraphy,oranabnormalultrasound.
Analysis
ThediagnosticaccuracyoftheF1+2assaywasmeasuredbycalculatingthesensitivity, specificity and the NPV for the usual cutoff value of 0.5nmol/L. Furthermore, we analyzed whether decreasing this cutoff to 0.3 and 0.4 or increasing it to 0.6 and 0.7nmol/Lwouldprovideabetterdiagnosticaccuracy.TodeterminetheoverallperformanceoftheF1+2testtheareaundertheROCcurve wascalculatedandcomparedtothatoftheDdimerassay.
Finally,weanalyzedwhetherF1+2couldbeofvalueinincreasingthepercentageof patients that could be withheld from additional imaging testing. Therefore the proportionofthisgroupwasestimatedandthesafetyofF1+2wasmeasuredbythe prevalenceofpulmonaryembolisminthesubgroupofpatientswithanabnormalD dimertestresult,andwithF1+2levelsbelowthe0.5nmol/Lcutofflevel.
Clinicalusefulnessofprothrombinfragment1+2
99 C HAPTER 7Table 1. Baseline clinical and demographic characteristics of the 510 study patients with clinically suspected pulmonary embolism
D-dimer and F1+2 available Characteristics All patients D-dimer and
F1+2 available
VTE – VTE +
Number of patients, (%) 510 373 (73%) 305 (82%) 68 (18%)
Female sex, n (%) 302 (59%) 214 (57%) 173 (57%) 41 (58%)
Mean age, years (range) 52 (19-91) 53 (19-91) 51 (20-91) 58 (19-88)
Malignancy, n (%) 84 (17%) 54 (15%) 42 (14%) 12 (17%)
Previous VTE, n (%) 55 (11%) 44 (12%) 34 (11%) 10 (14%)
R
ESULTS
During the study period between May 1999 and April 2001, a total 510 consecutive patientswithsuspectedpulmonaryembolismwereincluded.BothF1+2andDdimer results were available in 373 patients (73%). The baseline characteristics of these patientsareshowninTable1andaresimilartothecharacteristicsofthetotalcohort. Pulmonary embolism was ruled out in 41 patients (11%), based on a pretest probability of less than 20% of having the disease and a normal Ddimer. The remaining 332 patients underwent additional imaging, which confirmed pulmonary embolism in 68 patients (prevalence of pulmonary embolism 18%). Hence, for the analysis of F1+2 and Ddimer, 68 patients were considered to have pulmonary embolism,whereas305didnothavePE(Table1).
InpatientswithoutpulmonaryembolismtheF1+2valuesvariedbetween0.16nmol/L and8.19nmol/L,withameanof0.97nmol/L(SEM0.05)andamedianof0.7nmol/L.In those with pulmonary embolism F1+2 levels ranged between 0.33nmol/L and 8.22nmol/lwithameanof1.39nmol/l(SEM0.21)andamedianof1.25nmol/L.
Thesensitivity,specificityandNPVforthepresenceofpulmonaryembolismforfive differentcutofflevelsofF1+2aredepictedinTable2.Atthepredefinedcutoffvalue of0.5nmol/Lthesensitivitywas94%(95%CI8899%),thespecificity27%(95%CI22 32%) and NPV 96% (95%CI 9099%). As expected, the specificity increased using higher cutoff values, whereas the sensitivity and the NPV decreased. These characteristics were 96%, 45% and 98%, respectively, for the Ddimer assay (cutoff value0.5mg/L).
100
Table 2. Sensitivity, specificity and negative predictive value for F1+2 for different cut-offs.
F1+2 cut-off Sensitivity Specificity Negative predictive
value 0.3 nmol/L 100% 2% 100% 0.4 nmol/L 97% 12% 95% 0.5 nmol/L 94% 27% 96% 0.6 nmol/L 89% 37% 93% 0.7 nmol/L 79% 47% 91% TheROCcurveoftheF1+2assaywhenusedasthesoletestindiagnosingpulmonary embolismisdepictedinFigure1.Theareaunderthecurveis0.69(95%CI0.630.76). ForcomparisontheROCfortheDdimerassayisalsodepictedinFigure1.Thearea underthecurveis0.82(95%CI0.770.87),whichishigherthanforF1+2(p<0.05). 1,0 0,8 0,6 0,4 0,2 0,0 1 - Specificity 1,0 0,8 0,6 0,4 0,2 0,0 Sen sitivity
Figure 1. ROC curve of the F1+2 and D-dimer assay. The area under the curve with its 95% confidence intervals is 0.69 (0.63-0.76) for F1+2 and 0.82 (0.77-0.87) for D-dimer.
D-dimer F1+2 Source of the Curve
Clinicalusefulnessofprothrombinfragment1+2
101 C HAPTER 7To determine whether F1+2 could be used in a subgroup of patients, to increase the percentage of patients in whom pulmonary embolism could be safely excluded, the additionaldiagnosticvalueinpatientswithahighDdimerresultwasanalyzed. Ofthetotalstudypopulation,63%ofthepatientshadanabnormalDdimer.Twenty of these patients (8%) had an F1+2 below the 0.5nmol/L cutoff, and pulmonary embolismwasconfirmedinonepatient.Thesensitivity,specificityandNPVwith95% CIintervalsofF1+2inthissubgroupofpatientswithanabnormalDdimertherefore are99%(93100),10%(616)and95%(75100),respectively.
D
ISCUSSION
ThemainresultsofourstudyindicatethatF1+2isoflimitedvalueinthediagnostic workup of patients with suspected pulmonary embolism. Firstly, to obtain sensitivities approaching 100%, cutoff levels for F1+2 had to be set at 0.4nmol/L or lower. However, the associated specificity than becomes very low (Table 2; 212%), resulting in a marginal clinical utility, in terms of efficiently excluding the disease. Secondly, compared to the Ddimer test the area under the curve was inferior (0.69 and0.82,respectively,p<0.05)revealingthatoverthefullrangeofcutoffvaluesthe Ddimer test provides better diagnostic information. Finally, taking a normal F1+2 levelintoaccountinthesubgroupofpatientswithanabnormalDdimeraddedlittle clinical significance, since only 20 patients (8%) from this subgroup could be identified with pulmonary embolism in one of them. Therefore, including F1+2 in a diagnosticstrategynexttoDdimermeasurementappearsneitherefficientnorsafe.
We had expected more overlap in the diagnostic accuracy of the F1+2 and Ddimer testresults,sincetheyarebothreflectionsofcoagulationactivation.TheoreticallyF1+2 testsmayevenbemorespecific.ThereasonforthebetterperformanceofDdimeris notclear,butmaybeduetothelongerhalflifeofDdimer.
Somepotentiallimitationsofourstudyneedtobeaddressed.TestresultsofDdimer and F1+2 were available in 373 patients (73%) of the initial cohort of 510 patients. However,sincetheclinicalcharacteristicsofthesetwogroupsdidnotdiffer(Table1) itisunlikelythataselectionbiaswaspresent.Inthecurrentstudyclinicalprobability was estimated by the gestalt technique and ventilation perfusion scanning was the major imaging test. Currently these methods have been replaced by formal
102
quantitative clinical decision rules and spiral CT. Nevertheless, since this is a comparison of two blood tests in the same population we believe that our findings remain valid also for more modern diagnostic strategies. This is a retrospective analysiswithamoderatesamplesize.However,bloodwascollectedatbaselineand outcomes were assessed independently. Furthermore, the assay for F1+2 was done batchwiseandclassifiedwithoutknowledgeofthediagnosis.
Inconclusion,thediagnosticaccuracyofF1+2inpatientswithsuspectedpulmonary embolism is less efficient than the currently used Ddimer tests. Furthermore, the additional value of F1+2 is limited in the subgroup of patients with an abnormal D dimer.
Clinicalusefulnessofprothrombinfragment1+2
103 C HAPTER 7R
EFERENCELIST
1. van Belle A, Buller HR, Huisman MV et al. Effectiveness of managing suspected pulmonary
embolism using an algorithm combining clinical probability, D-dimer testing, and computed tomography. JAMA. 2006;295:172-179.
2. Ten Cate-Hoek AJ, Prins MH. Management studies using a combination of D-dimer test
result and clinical probability to rule out venous thromboembolism: a systematic review. J
Thromb Haemost. 2005;3:2465-2470.
3. Di Nisio M, Squizzato A, Rutjes AW, Buller HR, Zwinderman AH, Bossuyt PM. Diagnostic
accuracy of D-dimer test for exclusion of venous thromboembolism: a systematic review. J
Thromb Haemost. 2007;5:296-304.
4. Boneu B, Bes G, Pelzer H, Sie P, Boccalon H. D-Dimers, thrombin antithrombin III
complexes and prothrombin fragments 1+2: diagnostic value in clinically suspected deep vein thrombosis. Thromb Haemost. 1991;65:28-31.
5. Corradi A, Lazzaro F, Cofrancesco E, Cortellaro M, Ravasi F, Bertocchi F. Preoperative
plasma levels of prothrombin fragment 1 + 2 correlate with the risk of venous thrombosis after elective hip replacement. Acta Orthop Belg. 1999;65:39-43.
6. LaCapra S, Arkel YS, Ku DH, Gibson D, Lake C, Lam X. The use of thrombus precursor
protein, D-dimer, prothrombin fragment 1.2, and thrombin antithrombin in the exclusion of proximal deep vein thrombosis and pulmonary embolism. Blood Coagul Fibrinolysis. 2000;11:371-377.
7. Bozic M, Blinc A, Stegnar M. D-dimer, other markers of haemostasis activation and soluble
adhesion molecules in patients with different clinical probabilities of deep vein thrombosis.
Thromb Res. 2002;108:107-114.
8. Cofrancesco E, Cortellaro M, Corradi A, Ravasi F, Bertocchi F. Clinical utility of
prothrombin fragment 1+2, thrombin antithrombin III complexes and D-dimer measurements in the diagnosis of deep vein thrombosis following total hip replacement.
Thromb Haemost. 1998;79:509-510.
9. Ten Wolde M, Hagen PJ, Macgillavry MR et al. Non-invasive diagnostic work-up of patients
with clinically suspected pulmonary embolism; results of a management study. J Thromb
Haemost. 2004;2:1110-1117.
10. Ten Wolde M, Hagen PJ, Macgillavry MR et al. Non-invasive diagnostic work-up of patients with clinically suspected pulmonary embolism; results of a management study. J Thromb
Haemost. 2004;2:1110-1117.
