- Email: [email protected]
Exclusion of venous thromboembolism: evaluation of D-Dimer PLUS for the quantitative determination of D-dimer
Thrombosis Research (2005) 115, 381 — 386
intl.elsevierhealth.com/journals/thre
REGULAR ARTICLE
Exclusion of venous thromboembolism: evaluation of D-Dimer PLUS for the quantitative determination of D-dimer H.J. Vermeera,b,*, P. Ypmaa, M.J.L. van Strijenc, A.A. Muradina, F. Hudigb, R.W. Jansenb, P.W. Wijermansa, W.B.J. Gerritsa a
Department of Haematology, The Hague, The Netherlands Department of Clinical Chemistry, The Hague, The Netherlands c Department of Radiology of Haga Hospital, The Hague, The Netherlands b
Received 5 July 2004; received in revised form 20 August 2004; accepted 9 September 2004 Available online 27 October 2004
KEYWORDS Pulmonary embolism; Deep venous thrombosis; D-Dimer PLUS; Fibrin fibrinogen degradation products
Abstract The objective of this study was to evaluate if D-Dimer PLUS (Dade Behring, USA), a rapid fully automated assay, could be used as an initial screening test in the diagnosis of venous thromboembolism (VTE). Samples from 274 consecutive symptomatic patients with suspected pulmonary embolism (n=229; 79% outpatients, 21% inpatients), deep venous thrombosis (n=37; 84% outpatients, 16% inpatients) or suspected for both complications (n=8) were tested with this Ddimer assay with a Sysmex CA-1500 Coagulation Analyzer. Clinical probability for pulmonary embolism (PE) or deep venous thrombosis (DVT) was staged according to a pretest risk score proposed by Wells. Final diagnosis of PE and/or DVT was established by spiral-computed tomography of the pulmonary arteries or compression ultrasonography, respectively. PE was diagnosed in 13.5% of the patients, whereas DVT was confirmed in 17.7% of the patients. The optimal cutoff value for exclusion of venous thromboembolism was 130 Ag/l, and sensitivity, specificity and negative predictive value (NPV) were 95.0% (95% CI: 92.4 97.6), 30.4% (95% CI: 25.0 35.8) and 97.2% (95% CI: 95.2 99.2), respectively. In fact, two patient with PE were missed using D-Dimer PLUS; both cases were outpatients. In conclusion, this assay appears to be safe when implemented in an algorithm based on clinical assessment, D-dimer concentration, and radiological diagnostic techniques to stratify the risk for PE or DVT. However, higher
* Corresponding author. Leyenburg Hospital, Leyweg 275, PO Box 40551, 2504 LN The Hague, The Netherlands. Tel.: +31 70 3592269; fax: +31 70 359 2191. E-mail address: [email protected] (H.J. Vermeer). 0049-3848/$ - see front matter D 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.thromres.2004.09.005
382
H.J. Vermeer et al. sensitivities and negative predictive values were claimed in the scarce published reports for the D-Dimer PLUS assay than found in this study. D 2004 Elsevier Ltd. All rights reserved.
Venous thromboembolism is a frequent disease and potentially fatal in case of pulmonary embolism [1]. The exact incidence of pulmonary embolism is uncertain, hospital autopsy studies indicate that in a significant number of patients dying of pulmonary embolism, the diagnosis was not considered. When the diagnosis is made, effective therapy can be initiated and death associated with pulmonary embolism seems to become rare. However, an exception to this are patients who present with large emboli and initial hemodynamic impairment. Also, late mortality due to pulmonary hypertension as a result of high degree of obstruction of the pulmonary vasculature is a discernible problem. In spite of effective therapy the 3-month mortality is as high as 17% [2,3]. It is stated, still, that the majority of preventable deaths related to pulmonary embolism can be attributed to failure or missing of the diagnosis [4,5]. On the other hand, less than 30% of all patients presenting with signs and symptoms suggestive of pulmonary embolism apparently show embolism with the applied imaging technique [6,7]. Diagnosing pulmonary emboli therefore is a difficult task; clinical manifestations often are nonspecific [8]. Numerous studies have shown that D-dimer can be used for ruling out acute DVT or PE in combination with the use of a clinical decision rule. Indeed, clinical assessment on its own is insufficient to establish or exclude the diagnosis VTE, whereas a D-dimer assay also cannot be used as stand-alone test [9,10]. The sensitivity, specificity and negative predictive value vary depending on the type of D-dimer assay, but with the current rapid tests, both sensitivity and negative predictive value are usually high (mostly N95%) [11—14]. In fact, to safely exclude PE, the sensitivity and negative predictive value should approach 100%. In the current study, we describe the clinical evaluation of an immunoturbidimetric D-dimer assay: D-Dimer PLUS (Dade Behring) on the Sysmex CA-1500 analyzer; this rapid and automated test is particularly suited for emergency use. The question is: how accurate is this D-dimer assay in the diagnostic management of PE and/or DVT? To our knowledge, until now only very limited reports are published dealing with D-
Dimer PLUS results in terms of analytical and clinical performance [11]. A prospective study was started in our hospital to investigate the robustness of D-Dimer PLUS assay in a combination of a priori clinical probability test according to Wells et al. [15,16], diagnostic imaging procedures, and DDimer PLUS test in outpatients referred to our hospital with clinically suspected VTE or inpatients presumed to suffer from VTE. Besides this, also some hospitalised patients were involved in this study as relatively little is known about efficacy of D-dimer testing in this group of patients. Patients presenting with signs and symptoms raising the suspicion of pulmonary embolism or deep venous thrombosis, respectively, underwent spiral-computed tomography of the pulmonary vasculature or compression ultrasonography of the limb affected. Our hospital took part in multicentre clinical management studies to determine the effectiveness and safety of using spiral CT scanning as a primary diagnostic-imaging test [17—19]. For this reason, neither ventilation/perfusion scintigraphy nor pulmonary angiography was performed in the diagnostic work-up.
Patients and methods D-Dimer PLUS (Dade Behring, USA) is a latexenhanced turbidimetric test for the quantitative determination of cross-linked fibrin degradation products (D-Dimer) in human plasma. Polystyrene particles are covalently linked to a monoclonal antibody (DD5) to the cross-linkage region of crosslinked degradation products. D-Dimer PLUS was measured on a Sysmex CA-1500 System, a fully automated coagulation analyzer. D-dimer measurements were performed in platelet poor citrate plasma in duplicate. Intra-assay and inter-assay precision of the assay was determined according to NCCLS guidelines. D-Dimer PLUS assays in this study were performed and interpreted by independent operators without knowledge of the radiographic results. Method comparison was performed on 76 samples, with VIDAS D-Dimer assay (VIDAS D-
Evaluation of D-Dimer PLUS Table 1
383
Patient characteristics
Characteristic
Patients suspected for PE
Patients suspected for DVT
N Male Mean age Female Mean age Inpatients Outpatients Wells’ score (outpatients)b
229 71 60.6F17.1 years 158 54.9F20.7 years 47 182 low probability 79 moderate probability 74 high probability 14
37+8a 9+2a 55.7F19.4 years 28+6a 60.7F23.0 years 7 38 low probability 13 moderate probability 15 high probability 1
a These number of patients were suspected for both PE and DVT. b Wells’ scores are only calculated in cases of appropriate patient information.
Dimer New, BioMe ´rieux, Marcy l’Etoile, France) [20]. A prospective cohort study was performed from March 2003 to December 2003 in 274 consecutive patients referred to our hospital with suspected deep vein thrombosis (DVT) of the leg or suspected pulmonary embolism (PE) (220 outpatients, 54 hospitalised patients). Patient characteristics are shown in Table 1. Scores according to Wells et al. [15] were retrospectively blinded assigned by experienced clinicians derived from extended clinical assessment of the outpatient presenting to the emergency department (scores classified into low, moderate to high, and high). In this study, all patients underwent diagnostic imaging, either single-detector helical (spiral) computed tomography (suspected PE) or compression ultrasonography (suspected DVT). These techniques were at the same time the confirmational tests on the understanding that the CT scan was described using a specific pulmonary embolism protocol by a trained radiologist. The compression ultrasonography was repeated after 4 and 7 days in case of ongoing symptoms. To derive the optimal cut-off value, a receiver operator characteristic (ROC) curve was constructed by plotting sensitivity (true positive fraction) versus 1 specificity (false positive fraction), and area under the curve was calculated (AUC) using Analyze-it Software for Microsoft Excel (Leeds, UK). Sensitivity, negative predictive value (NPV) and specificity of D-Dimer PLUS was calculated using cross-table procedures in relation to the results of the clinical outcome diagnosed by spiral CT or duplex ultrasonography in case of pulmonary embolism or deep venous thrombosis, respectively. Correlation between groups were statistically evaluated using Wilcoxon rank sum test for unpaired
results or Pearson R correlation for paired results (SPSS/PC5).
Results An ROC curve was made displaying sensitivity and specificity for D-Dimer PLUS at different cut-off levels having an AUC of 0.75 (95% CI: 0.67—0.83; ROC curve not shown). From this ROC curve, a most appropriate cut-off value of 130 Ag/l for the management of VTE could be derived. This decision threshold is in concordance with recommendations of the manufacturer. Total CV of the D-Dimer PLUS assay around the decision level of 130 Ag/l was relatively high: 11.1%. Method comparison versus VIDAS, based on 76 patients, gave a correlation of 0.83 (Pearson correlation coefficient (R); 95% CI: 0.74—0.89). Results given in Table 2 combined with final diagnosis for these 76 patients showed that specificity was somewhat higher for D-Dimer PLUS compared to VIDAS (data not shown). There was no significant difference between Ddimer concentrations of patients diagnosed with PE or DVT (Wilcoxon rank sum test: W=121, p 0.917; hypothesis: population mean is different for PE and DVT). Therefore, in Table 3 the performance of D-Dimer PLUS in terms of sensitivity, specificity and negative predictive value is calculated for venous thromboembolism (VTE) without making difference between PE and DVT. Clinical sensitivities and negative predictive values for inpatients were 100% since both patients with PE and missed by D-Dimer PLUS were outpatients. Interestingly, the specificity was lower for inpatients [23.5% (95% CI: 12.7—34.3)] than outpatients [32.1% (95% CI: 26.0—38.2)]. The lower specificity for inpatients can be explained by the generally higher coagulation state of hospitalised patients due to comorbidity. In 196 (74%) of the 266 patients, the D-dimer test was abnormal (N130 Ag/l). A positive diagnosis of VTE was assessed for 31/229 (13.5%) of patients with suspected PE and 8/45 (17.7%) with suspected DVT. The clinical outcome was not significantly different according to sex or age (Wilcoxon rank Table 2 PLUS
Method comparison, VIDAS versus D-Dimer
D-Dimer PLUS
VIDAS Positive
Negative
Positive Negative
46 8
4 18
384
H.J. Vermeer et al.
Table 3
Performance of D-Dimer PLUS for exclusion of VTEa
Decision threshold (Ag/l)
Sensitivity (%) (95% CI)
Specificity (%) (95% CI)
NPVb (95% CI)
PPVc (95% CI)
100 130 160 190 220
95.0 95.0 92.3 92.3 87.2
22.5 30.4 39.6 44.5 52.0
96.3 97.2 96.8 97.1 95.9
17.4 19.0 20.8 22.2 23.8
a b c
(92.4 (92.4 (89.1 (89.1 (83.2
97.6) 97.6) 95.5) 95.5) 91.2)
(17.6 (25.0 (33.8 (38.6 (46.1
27.4) 35.8) 45.4) 50.4) 57.9)
(94.1 (95.2 (94.7 (95.1 (93.6
98.5) 99.2) 98.9) 99.1) 98.2)
(12.9 (14.4 (16.0 (17.3 (18.8
21.9) 23.6) 25.6) 27.1) 28.8)
Venous thromboembolism. Negative predictive value. Positive predictive value.
sum test, data not shown). As can be seen in Fig. 1, two patients suspected for PE were missed by the D-dimer assay; both samples had D-dimer concentrations b50 Ag/l. For one of these patients, the result on the VIDAS D-Dimer assay was also below the cut-off value. All patients with PE or DVT and positive D-Dimer PLUS results had also positive VIDAS results. Wells’ scores for outpatients were almost concordant with the final diagnosis: most patients with confirmed VTE had moderate or high pretest probability for VTE. However, one patient with diagnosed PE exerted a low score for PE, whereas the D-dimer concentration was above the cut-off level. Alternatively, no patients with low pretest probabilities and negative D-dimer results (thus below cut-off value) had a confirmed diagnosis of PE or DVT. For patients with confirmed PE and Ddimer results below cut-off (n=2), Wells’ scores predicted a moderate to high probability for PE. From patients with a moderate to high pretest probability for PE, in 24.1% (n=88) of the cases the D-dimer was negative (i.e. below cut-off value). In case of patients having a moderate to high risk
score for DVT, D-dimer was negative in 40% (n=16) of the patients.
Discussion The demands on a D-dimer assay in the diagnostic management of VTE are very high. Results of Ddimer assays of different manufacturers are not interchangeable and commercially available kits differ in their reaction towards D-dimer fragments [21]. Furthermore, there is still no reference method for D-dimer detection. The magnitude of the CV for D-Dimer PLUS around the cut-off value is over 10% and thus higher than published data (wherein CVs ranged from 2.6% to 5.2%), and measured concentrations in the cut-off range can result in false-negative results. In our study, 20 (7.5%) patients had results in the range of 130 Ag/lF11% (i.e. range 116—144 Ag/l) and we recommend duplicate measurements when test results fall into this range. Our clinical validation data show that a cut-off value of 130 Ag/l for the Sysmex CA-1500 system
Figure 1 Individual D-Dimer PLUS results for patients with and without pulmonary embolism (PE) or deep venous thrombosis (DVT). The broken line represents the decision threshold (=cut-off value).
Evaluation of D-Dimer PLUS seems adequate. So, we confirmed data of Reber et al. [11] suggesting a cut-off value of 130 Ag/l for this D-dimer assay on a BCT Analyzer, having a sensitivity, specificity and negative predictive value of 96.8% (95% CI: 91.0 99.3), 45.2% (95% CI: 38.5 51.8), and 97.0% (95% CI: 91.6 99.4), respectively. However, in our study lower performances were found due to more false negative results (n=2): 95.0% (95% CI: 92.4 97.6), 30.4% (95% CI: 25.0 35.8), and 97.2% (95% CI: 95.2 99.2), respectively, as compared to Reber et al. [11]. The reason for the missed PE is not clear; however, for one patient the negative D-dimer was confirmed by VIDAS D-dimer measurement. Unfortunately, no material was left for the other patient to test this patient with the VIDAS assay. Besides this, both patients had moderate to high Wells’ scores. Thus, in an algorithm using D-dimer for evaluation of suspected PE, the management of these patients with discordant results should have implicated further evaluation with CT scan. However, it must be emphasized here that higher sensitivities and negative predictive values were found for other D-dimer assays [12,13,22—24]. We have found that the combination of a D-Dimer PLUS result below the cut-off and a low pretest clinical probability can be successfully used in an algorithm for evaluation of suspected PE or DVT. In this study, a diagnosis based on D-dimer measurement and low clinical risk could have been resulted in a 12% decrease of performed lung scans. According to the algorithm to be implemented, in case of a negative D-dimer and low pretest score no further diagnostic evaluation will be done, and anticoagulant therapy will be withheld. In all other situations, further diagnostic management will be necessary. Moreover, the strength of a combination of negative D-dimer and low pretest score is also illustrated in this study as one patient with a low score and a positive D-dimer was finally diagnosed to have PE. On the other hand, as stated before, two patients with negative D-dimer values had moderate to high pretest probability scores. So, such patients are not missed if the algorithm is implemented. We recommend more studies for aberrant patient groups (e.g. pregnancy, coronary disease, cancer, liver disease, inflammation) to evaluate if the scope of D-dimer in the implementation in diagnostic algorithms can be broadened. In conclusion, D-Dimer PLUS in combination with preclinical scores provides a useful tool for the clinician to exclude safely PE or DVT. However, the clinical sensitivity, specificity and negative predictive value in the exclusion of VTE found here are lower than described for most other currently used D-dimer assays. So, it would be wise to confirm
385 these parameters for D-Dimer PLUS in other studies in order to collect more data from patients suspected for VTE.
Acknowledgements The authors are indebted to K.C. van de Noort for collecting reports of outpatients presenting in the emergency department. The continuous support of W. van Spronsen-Hatzmann is greatly appreciated. The authors also would like to thank all laboratory and radiological technicians for their contributions to the work described here.
References [1] Kearon C, Hirsh J. The diagnosis of pulmonary embolism. Haemostasis 1995;25:72 — 87. [2] Hull RD, Hirsh J, Carter CJ, Jay RM, Dodd PE, Ockelford PA, et al. Pulmonary angiography, ventilation lung scanning, and venography for clinically suspected pulmonary embolism with abnormal perfusion lung scan. Ann Intern Med 1983;98:891 — 9. [3] Stein PD, Hull RD, Saltzman HA, Pineo G. Strategy for diagnosis of patients with suspected acute pulmonary embolism. Chest 1993;103:1553 — 9. [4] Fedullo PF, Tapson VF. The evaluation of suspected pulmonary embolism. NEJM 2003;349:1247 — 56. [5] Wells PS, Hirsh J, Anderson DR, Lensing AWA, Foster G, Kearon C, et al. Accuracy of clinical assessment of deepvein thrombosis. Lancet 1995;345:1326 — 9. [6] Anand SS, Wells PS, Hunt D, Brill-Edwards P, Cook D, Ginsberg JS. Does this patient have deep vein thrombosis? JAMA 1998;279:1094 — 9. [7] Wells PS, Anderson DR, Bormanis J, Guy F, Mitchell M, Gray L, et al. Value of assessment of pretest probability of deepvein thrombosis in clinical management. Lancet 1997; 350:1795 — 8. [8] Wells PS, Hirsh J, Anderson DR, Lensing AW, Foster G, Kearon C, et al. A simple clinical model for the diagnosis of deep-vein thrombosis combined with impedance plethysmography: potential for an improvement in the diagnostic process. J Intern Med 1998;243:15 — 23. [9] Reber G, Bounameaux H, Perrier A, et al. Evaluation of advanced D-Dimer assay for the exclusion of venous thromboembolism. Thromb Res 2002;107:197 — 200. [10] Schutgens REG, Haas FJLM, Gerritsen WBM, Van der Horst F, Nieuwenhuis HK, Biesma DH. The usefulness of five D-dimer assays in the exclusion of deep venous thrombosis. J Thromb Haemost 2003;1:976 — 81. [11] Gosselin RC, Owings JT, Kehoe J, Anderson JT, Dwyre DM, Jacoby RC, et al. Comparison of six D-dimer methods in patients suspected of deep vein thrombosis. Blood Coagul Fibrinolysis 2003;14:545 — 50. [12] Maskell NA, Butland RJA. Comparison of a quantitative latex and a quantitative ELISA plasma D-dimer assay in the exclusion of segmental and subsegmental pulmonary embolism. Thromb Haemost 2001;86:1580 — 2. [13] Anderson Jr FA, Wheeler HB, Goldberg RJ, Hosmer DW, Patwardhan NA, Jovanovic B, et al. A population-based
386
[14]
[15] [16]
[17]
[18]
[19]
H.J. Vermeer et al. perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. Arch Intern Med 1991;151:933 — 8. Grifioni S, Olivotto I, Cecchini P, Pieralli F, Carnaili A, Santoro G, et al. Short-term clinical outcome of patients with acute pulmonary embolism, normal blood pressure, and echocardiographic right ventricular dysfunction. Circulation 2000;101:2817 — 22. Goldhaber SZ. Treatment of pulmonary thromboembolism. Intern Med 1999;38:620 — 5. Sandler DA, Martin JF. Autopsy proven pulmonary embolism in hospital patients: are we detecting enough deep vein thrombosis? J R Soc Med 1989;82:203 — 5. Value of the ventilation/perfusion scan in acute pulmonary embolism. Results of the prospective investigation of pulmonary embolism diagnosis (PIOPED). The PIOPED investigators. JAMA 1990;263:2753 — 9. Van Strijen MJL, De Monye ´ W, Kieft GJ, Pattynama PMT, Huisman MV, Smith SJ, et al. Diagnosis of pulmonary embolism with spiral CT as a second procedure following scintigraphy. Eur Radiol 2003;13:1501 — 7. Van Strijen MJL, De Monye ´ W, Schiereck J, Kieft GJ, Prins MV, Huisman MV, et al. Single detector CT as the primary
[20]
[21]
[22]
[23]
[24]
diagnostic test in suspected pulmonary embolism: a multicentre clinical management study in 510 patients. Ann Intern Med 2003;138:307 — 14. De Moerloose P, Desmarais S, Bounameaux H, Reber G, Perrier A, Dupuy G, et al. Contribution of a new, rapid, individual and quantitative automated D-dimer ELISA to exclude pulmonary embolism. Thromb Haemost 1996;75: 11 — 3. Dempfle C-E, Zips S, Ergu ¨l H, Heene DL, the FACT study group. The fibrin assay comparison trial (FACT). Evaluation of 23 quantitative D-dimer assays as basis for the development of D-dimer calibrators. Thromb Haemost 2001;85: 671 — 8. Van der Graaf F, Van den Borne H, Van der Kolk M, De Wild GWT, Janssen GWT, Van Uum SHM. Exclusion of deep venous thrombosis with D-dimer testing. Thromb Haemost 2000; 83:191 — 8. Gardiner C, Mackie IJ, Machin SJ. Evaluation of a new automated latex particle immunoassay for D-dimer: IL test D-dimer. Lab Hematol 2000;6:147 — 50. Bozˇicˇ M, Stegnar M. Validation of an automated immunoturbidimetric assay for measurement of plasma D-dimer. Clin Chem Lab Med 2003;41:958 — 62.
intl.elsevierhealth.com/journals/thre
REGULAR ARTICLE
Exclusion of venous thromboembolism: evaluation of D-Dimer PLUS for the quantitative determination of D-dimer H.J. Vermeera,b,*, P. Ypmaa, M.J.L. van Strijenc, A.A. Muradina, F. Hudigb, R.W. Jansenb, P.W. Wijermansa, W.B.J. Gerritsa a
Department of Haematology, The Hague, The Netherlands Department of Clinical Chemistry, The Hague, The Netherlands c Department of Radiology of Haga Hospital, The Hague, The Netherlands b
Received 5 July 2004; received in revised form 20 August 2004; accepted 9 September 2004 Available online 27 October 2004
KEYWORDS Pulmonary embolism; Deep venous thrombosis; D-Dimer PLUS; Fibrin fibrinogen degradation products
Abstract The objective of this study was to evaluate if D-Dimer PLUS (Dade Behring, USA), a rapid fully automated assay, could be used as an initial screening test in the diagnosis of venous thromboembolism (VTE). Samples from 274 consecutive symptomatic patients with suspected pulmonary embolism (n=229; 79% outpatients, 21% inpatients), deep venous thrombosis (n=37; 84% outpatients, 16% inpatients) or suspected for both complications (n=8) were tested with this Ddimer assay with a Sysmex CA-1500 Coagulation Analyzer. Clinical probability for pulmonary embolism (PE) or deep venous thrombosis (DVT) was staged according to a pretest risk score proposed by Wells. Final diagnosis of PE and/or DVT was established by spiral-computed tomography of the pulmonary arteries or compression ultrasonography, respectively. PE was diagnosed in 13.5% of the patients, whereas DVT was confirmed in 17.7% of the patients. The optimal cutoff value for exclusion of venous thromboembolism was 130 Ag/l, and sensitivity, specificity and negative predictive value (NPV) were 95.0% (95% CI: 92.4 97.6), 30.4% (95% CI: 25.0 35.8) and 97.2% (95% CI: 95.2 99.2), respectively. In fact, two patient with PE were missed using D-Dimer PLUS; both cases were outpatients. In conclusion, this assay appears to be safe when implemented in an algorithm based on clinical assessment, D-dimer concentration, and radiological diagnostic techniques to stratify the risk for PE or DVT. However, higher
* Corresponding author. Leyenburg Hospital, Leyweg 275, PO Box 40551, 2504 LN The Hague, The Netherlands. Tel.: +31 70 3592269; fax: +31 70 359 2191. E-mail address: [email protected] (H.J. Vermeer). 0049-3848/$ - see front matter D 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.thromres.2004.09.005
382
H.J. Vermeer et al. sensitivities and negative predictive values were claimed in the scarce published reports for the D-Dimer PLUS assay than found in this study. D 2004 Elsevier Ltd. All rights reserved.
Venous thromboembolism is a frequent disease and potentially fatal in case of pulmonary embolism [1]. The exact incidence of pulmonary embolism is uncertain, hospital autopsy studies indicate that in a significant number of patients dying of pulmonary embolism, the diagnosis was not considered. When the diagnosis is made, effective therapy can be initiated and death associated with pulmonary embolism seems to become rare. However, an exception to this are patients who present with large emboli and initial hemodynamic impairment. Also, late mortality due to pulmonary hypertension as a result of high degree of obstruction of the pulmonary vasculature is a discernible problem. In spite of effective therapy the 3-month mortality is as high as 17% [2,3]. It is stated, still, that the majority of preventable deaths related to pulmonary embolism can be attributed to failure or missing of the diagnosis [4,5]. On the other hand, less than 30% of all patients presenting with signs and symptoms suggestive of pulmonary embolism apparently show embolism with the applied imaging technique [6,7]. Diagnosing pulmonary emboli therefore is a difficult task; clinical manifestations often are nonspecific [8]. Numerous studies have shown that D-dimer can be used for ruling out acute DVT or PE in combination with the use of a clinical decision rule. Indeed, clinical assessment on its own is insufficient to establish or exclude the diagnosis VTE, whereas a D-dimer assay also cannot be used as stand-alone test [9,10]. The sensitivity, specificity and negative predictive value vary depending on the type of D-dimer assay, but with the current rapid tests, both sensitivity and negative predictive value are usually high (mostly N95%) [11—14]. In fact, to safely exclude PE, the sensitivity and negative predictive value should approach 100%. In the current study, we describe the clinical evaluation of an immunoturbidimetric D-dimer assay: D-Dimer PLUS (Dade Behring) on the Sysmex CA-1500 analyzer; this rapid and automated test is particularly suited for emergency use. The question is: how accurate is this D-dimer assay in the diagnostic management of PE and/or DVT? To our knowledge, until now only very limited reports are published dealing with D-
Dimer PLUS results in terms of analytical and clinical performance [11]. A prospective study was started in our hospital to investigate the robustness of D-Dimer PLUS assay in a combination of a priori clinical probability test according to Wells et al. [15,16], diagnostic imaging procedures, and DDimer PLUS test in outpatients referred to our hospital with clinically suspected VTE or inpatients presumed to suffer from VTE. Besides this, also some hospitalised patients were involved in this study as relatively little is known about efficacy of D-dimer testing in this group of patients. Patients presenting with signs and symptoms raising the suspicion of pulmonary embolism or deep venous thrombosis, respectively, underwent spiral-computed tomography of the pulmonary vasculature or compression ultrasonography of the limb affected. Our hospital took part in multicentre clinical management studies to determine the effectiveness and safety of using spiral CT scanning as a primary diagnostic-imaging test [17—19]. For this reason, neither ventilation/perfusion scintigraphy nor pulmonary angiography was performed in the diagnostic work-up.
Patients and methods D-Dimer PLUS (Dade Behring, USA) is a latexenhanced turbidimetric test for the quantitative determination of cross-linked fibrin degradation products (D-Dimer) in human plasma. Polystyrene particles are covalently linked to a monoclonal antibody (DD5) to the cross-linkage region of crosslinked degradation products. D-Dimer PLUS was measured on a Sysmex CA-1500 System, a fully automated coagulation analyzer. D-dimer measurements were performed in platelet poor citrate plasma in duplicate. Intra-assay and inter-assay precision of the assay was determined according to NCCLS guidelines. D-Dimer PLUS assays in this study were performed and interpreted by independent operators without knowledge of the radiographic results. Method comparison was performed on 76 samples, with VIDAS D-Dimer assay (VIDAS D-
Evaluation of D-Dimer PLUS Table 1
383
Patient characteristics
Characteristic
Patients suspected for PE
Patients suspected for DVT
N Male Mean age Female Mean age Inpatients Outpatients Wells’ score (outpatients)b
229 71 60.6F17.1 years 158 54.9F20.7 years 47 182 low probability 79 moderate probability 74 high probability 14
37+8a 9+2a 55.7F19.4 years 28+6a 60.7F23.0 years 7 38 low probability 13 moderate probability 15 high probability 1
a These number of patients were suspected for both PE and DVT. b Wells’ scores are only calculated in cases of appropriate patient information.
Dimer New, BioMe ´rieux, Marcy l’Etoile, France) [20]. A prospective cohort study was performed from March 2003 to December 2003 in 274 consecutive patients referred to our hospital with suspected deep vein thrombosis (DVT) of the leg or suspected pulmonary embolism (PE) (220 outpatients, 54 hospitalised patients). Patient characteristics are shown in Table 1. Scores according to Wells et al. [15] were retrospectively blinded assigned by experienced clinicians derived from extended clinical assessment of the outpatient presenting to the emergency department (scores classified into low, moderate to high, and high). In this study, all patients underwent diagnostic imaging, either single-detector helical (spiral) computed tomography (suspected PE) or compression ultrasonography (suspected DVT). These techniques were at the same time the confirmational tests on the understanding that the CT scan was described using a specific pulmonary embolism protocol by a trained radiologist. The compression ultrasonography was repeated after 4 and 7 days in case of ongoing symptoms. To derive the optimal cut-off value, a receiver operator characteristic (ROC) curve was constructed by plotting sensitivity (true positive fraction) versus 1 specificity (false positive fraction), and area under the curve was calculated (AUC) using Analyze-it Software for Microsoft Excel (Leeds, UK). Sensitivity, negative predictive value (NPV) and specificity of D-Dimer PLUS was calculated using cross-table procedures in relation to the results of the clinical outcome diagnosed by spiral CT or duplex ultrasonography in case of pulmonary embolism or deep venous thrombosis, respectively. Correlation between groups were statistically evaluated using Wilcoxon rank sum test for unpaired
results or Pearson R correlation for paired results (SPSS/PC5).
Results An ROC curve was made displaying sensitivity and specificity for D-Dimer PLUS at different cut-off levels having an AUC of 0.75 (95% CI: 0.67—0.83; ROC curve not shown). From this ROC curve, a most appropriate cut-off value of 130 Ag/l for the management of VTE could be derived. This decision threshold is in concordance with recommendations of the manufacturer. Total CV of the D-Dimer PLUS assay around the decision level of 130 Ag/l was relatively high: 11.1%. Method comparison versus VIDAS, based on 76 patients, gave a correlation of 0.83 (Pearson correlation coefficient (R); 95% CI: 0.74—0.89). Results given in Table 2 combined with final diagnosis for these 76 patients showed that specificity was somewhat higher for D-Dimer PLUS compared to VIDAS (data not shown). There was no significant difference between Ddimer concentrations of patients diagnosed with PE or DVT (Wilcoxon rank sum test: W=121, p 0.917; hypothesis: population mean is different for PE and DVT). Therefore, in Table 3 the performance of D-Dimer PLUS in terms of sensitivity, specificity and negative predictive value is calculated for venous thromboembolism (VTE) without making difference between PE and DVT. Clinical sensitivities and negative predictive values for inpatients were 100% since both patients with PE and missed by D-Dimer PLUS were outpatients. Interestingly, the specificity was lower for inpatients [23.5% (95% CI: 12.7—34.3)] than outpatients [32.1% (95% CI: 26.0—38.2)]. The lower specificity for inpatients can be explained by the generally higher coagulation state of hospitalised patients due to comorbidity. In 196 (74%) of the 266 patients, the D-dimer test was abnormal (N130 Ag/l). A positive diagnosis of VTE was assessed for 31/229 (13.5%) of patients with suspected PE and 8/45 (17.7%) with suspected DVT. The clinical outcome was not significantly different according to sex or age (Wilcoxon rank Table 2 PLUS
Method comparison, VIDAS versus D-Dimer
D-Dimer PLUS
VIDAS Positive
Negative
Positive Negative
46 8
4 18
384
H.J. Vermeer et al.
Table 3
Performance of D-Dimer PLUS for exclusion of VTEa
Decision threshold (Ag/l)
Sensitivity (%) (95% CI)
Specificity (%) (95% CI)
NPVb (95% CI)
PPVc (95% CI)
100 130 160 190 220
95.0 95.0 92.3 92.3 87.2
22.5 30.4 39.6 44.5 52.0
96.3 97.2 96.8 97.1 95.9
17.4 19.0 20.8 22.2 23.8
a b c
(92.4 (92.4 (89.1 (89.1 (83.2
97.6) 97.6) 95.5) 95.5) 91.2)
(17.6 (25.0 (33.8 (38.6 (46.1
27.4) 35.8) 45.4) 50.4) 57.9)
(94.1 (95.2 (94.7 (95.1 (93.6
98.5) 99.2) 98.9) 99.1) 98.2)
(12.9 (14.4 (16.0 (17.3 (18.8
21.9) 23.6) 25.6) 27.1) 28.8)
Venous thromboembolism. Negative predictive value. Positive predictive value.
sum test, data not shown). As can be seen in Fig. 1, two patients suspected for PE were missed by the D-dimer assay; both samples had D-dimer concentrations b50 Ag/l. For one of these patients, the result on the VIDAS D-Dimer assay was also below the cut-off value. All patients with PE or DVT and positive D-Dimer PLUS results had also positive VIDAS results. Wells’ scores for outpatients were almost concordant with the final diagnosis: most patients with confirmed VTE had moderate or high pretest probability for VTE. However, one patient with diagnosed PE exerted a low score for PE, whereas the D-dimer concentration was above the cut-off level. Alternatively, no patients with low pretest probabilities and negative D-dimer results (thus below cut-off value) had a confirmed diagnosis of PE or DVT. For patients with confirmed PE and Ddimer results below cut-off (n=2), Wells’ scores predicted a moderate to high probability for PE. From patients with a moderate to high pretest probability for PE, in 24.1% (n=88) of the cases the D-dimer was negative (i.e. below cut-off value). In case of patients having a moderate to high risk
score for DVT, D-dimer was negative in 40% (n=16) of the patients.
Discussion The demands on a D-dimer assay in the diagnostic management of VTE are very high. Results of Ddimer assays of different manufacturers are not interchangeable and commercially available kits differ in their reaction towards D-dimer fragments [21]. Furthermore, there is still no reference method for D-dimer detection. The magnitude of the CV for D-Dimer PLUS around the cut-off value is over 10% and thus higher than published data (wherein CVs ranged from 2.6% to 5.2%), and measured concentrations in the cut-off range can result in false-negative results. In our study, 20 (7.5%) patients had results in the range of 130 Ag/lF11% (i.e. range 116—144 Ag/l) and we recommend duplicate measurements when test results fall into this range. Our clinical validation data show that a cut-off value of 130 Ag/l for the Sysmex CA-1500 system
Figure 1 Individual D-Dimer PLUS results for patients with and without pulmonary embolism (PE) or deep venous thrombosis (DVT). The broken line represents the decision threshold (=cut-off value).
Evaluation of D-Dimer PLUS seems adequate. So, we confirmed data of Reber et al. [11] suggesting a cut-off value of 130 Ag/l for this D-dimer assay on a BCT Analyzer, having a sensitivity, specificity and negative predictive value of 96.8% (95% CI: 91.0 99.3), 45.2% (95% CI: 38.5 51.8), and 97.0% (95% CI: 91.6 99.4), respectively. However, in our study lower performances were found due to more false negative results (n=2): 95.0% (95% CI: 92.4 97.6), 30.4% (95% CI: 25.0 35.8), and 97.2% (95% CI: 95.2 99.2), respectively, as compared to Reber et al. [11]. The reason for the missed PE is not clear; however, for one patient the negative D-dimer was confirmed by VIDAS D-dimer measurement. Unfortunately, no material was left for the other patient to test this patient with the VIDAS assay. Besides this, both patients had moderate to high Wells’ scores. Thus, in an algorithm using D-dimer for evaluation of suspected PE, the management of these patients with discordant results should have implicated further evaluation with CT scan. However, it must be emphasized here that higher sensitivities and negative predictive values were found for other D-dimer assays [12,13,22—24]. We have found that the combination of a D-Dimer PLUS result below the cut-off and a low pretest clinical probability can be successfully used in an algorithm for evaluation of suspected PE or DVT. In this study, a diagnosis based on D-dimer measurement and low clinical risk could have been resulted in a 12% decrease of performed lung scans. According to the algorithm to be implemented, in case of a negative D-dimer and low pretest score no further diagnostic evaluation will be done, and anticoagulant therapy will be withheld. In all other situations, further diagnostic management will be necessary. Moreover, the strength of a combination of negative D-dimer and low pretest score is also illustrated in this study as one patient with a low score and a positive D-dimer was finally diagnosed to have PE. On the other hand, as stated before, two patients with negative D-dimer values had moderate to high pretest probability scores. So, such patients are not missed if the algorithm is implemented. We recommend more studies for aberrant patient groups (e.g. pregnancy, coronary disease, cancer, liver disease, inflammation) to evaluate if the scope of D-dimer in the implementation in diagnostic algorithms can be broadened. In conclusion, D-Dimer PLUS in combination with preclinical scores provides a useful tool for the clinician to exclude safely PE or DVT. However, the clinical sensitivity, specificity and negative predictive value in the exclusion of VTE found here are lower than described for most other currently used D-dimer assays. So, it would be wise to confirm
385 these parameters for D-Dimer PLUS in other studies in order to collect more data from patients suspected for VTE.
Acknowledgements The authors are indebted to K.C. van de Noort for collecting reports of outpatients presenting in the emergency department. The continuous support of W. van Spronsen-Hatzmann is greatly appreciated. The authors also would like to thank all laboratory and radiological technicians for their contributions to the work described here.
References [1] Kearon C, Hirsh J. The diagnosis of pulmonary embolism. Haemostasis 1995;25:72 — 87. [2] Hull RD, Hirsh J, Carter CJ, Jay RM, Dodd PE, Ockelford PA, et al. Pulmonary angiography, ventilation lung scanning, and venography for clinically suspected pulmonary embolism with abnormal perfusion lung scan. Ann Intern Med 1983;98:891 — 9. [3] Stein PD, Hull RD, Saltzman HA, Pineo G. Strategy for diagnosis of patients with suspected acute pulmonary embolism. Chest 1993;103:1553 — 9. [4] Fedullo PF, Tapson VF. The evaluation of suspected pulmonary embolism. NEJM 2003;349:1247 — 56. [5] Wells PS, Hirsh J, Anderson DR, Lensing AWA, Foster G, Kearon C, et al. Accuracy of clinical assessment of deepvein thrombosis. Lancet 1995;345:1326 — 9. [6] Anand SS, Wells PS, Hunt D, Brill-Edwards P, Cook D, Ginsberg JS. Does this patient have deep vein thrombosis? JAMA 1998;279:1094 — 9. [7] Wells PS, Anderson DR, Bormanis J, Guy F, Mitchell M, Gray L, et al. Value of assessment of pretest probability of deepvein thrombosis in clinical management. Lancet 1997; 350:1795 — 8. [8] Wells PS, Hirsh J, Anderson DR, Lensing AW, Foster G, Kearon C, et al. A simple clinical model for the diagnosis of deep-vein thrombosis combined with impedance plethysmography: potential for an improvement in the diagnostic process. J Intern Med 1998;243:15 — 23. [9] Reber G, Bounameaux H, Perrier A, et al. Evaluation of advanced D-Dimer assay for the exclusion of venous thromboembolism. Thromb Res 2002;107:197 — 200. [10] Schutgens REG, Haas FJLM, Gerritsen WBM, Van der Horst F, Nieuwenhuis HK, Biesma DH. The usefulness of five D-dimer assays in the exclusion of deep venous thrombosis. J Thromb Haemost 2003;1:976 — 81. [11] Gosselin RC, Owings JT, Kehoe J, Anderson JT, Dwyre DM, Jacoby RC, et al. Comparison of six D-dimer methods in patients suspected of deep vein thrombosis. Blood Coagul Fibrinolysis 2003;14:545 — 50. [12] Maskell NA, Butland RJA. Comparison of a quantitative latex and a quantitative ELISA plasma D-dimer assay in the exclusion of segmental and subsegmental pulmonary embolism. Thromb Haemost 2001;86:1580 — 2. [13] Anderson Jr FA, Wheeler HB, Goldberg RJ, Hosmer DW, Patwardhan NA, Jovanovic B, et al. A population-based
386
[14]
[15] [16]
[17]
[18]
[19]
H.J. Vermeer et al. perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. Arch Intern Med 1991;151:933 — 8. Grifioni S, Olivotto I, Cecchini P, Pieralli F, Carnaili A, Santoro G, et al. Short-term clinical outcome of patients with acute pulmonary embolism, normal blood pressure, and echocardiographic right ventricular dysfunction. Circulation 2000;101:2817 — 22. Goldhaber SZ. Treatment of pulmonary thromboembolism. Intern Med 1999;38:620 — 5. Sandler DA, Martin JF. Autopsy proven pulmonary embolism in hospital patients: are we detecting enough deep vein thrombosis? J R Soc Med 1989;82:203 — 5. Value of the ventilation/perfusion scan in acute pulmonary embolism. Results of the prospective investigation of pulmonary embolism diagnosis (PIOPED). The PIOPED investigators. JAMA 1990;263:2753 — 9. Van Strijen MJL, De Monye ´ W, Kieft GJ, Pattynama PMT, Huisman MV, Smith SJ, et al. Diagnosis of pulmonary embolism with spiral CT as a second procedure following scintigraphy. Eur Radiol 2003;13:1501 — 7. Van Strijen MJL, De Monye ´ W, Schiereck J, Kieft GJ, Prins MV, Huisman MV, et al. Single detector CT as the primary
[20]
[21]
[22]
[23]
[24]
diagnostic test in suspected pulmonary embolism: a multicentre clinical management study in 510 patients. Ann Intern Med 2003;138:307 — 14. De Moerloose P, Desmarais S, Bounameaux H, Reber G, Perrier A, Dupuy G, et al. Contribution of a new, rapid, individual and quantitative automated D-dimer ELISA to exclude pulmonary embolism. Thromb Haemost 1996;75: 11 — 3. Dempfle C-E, Zips S, Ergu ¨l H, Heene DL, the FACT study group. The fibrin assay comparison trial (FACT). Evaluation of 23 quantitative D-dimer assays as basis for the development of D-dimer calibrators. Thromb Haemost 2001;85: 671 — 8. Van der Graaf F, Van den Borne H, Van der Kolk M, De Wild GWT, Janssen GWT, Van Uum SHM. Exclusion of deep venous thrombosis with D-dimer testing. Thromb Haemost 2000; 83:191 — 8. Gardiner C, Mackie IJ, Machin SJ. Evaluation of a new automated latex particle immunoassay for D-dimer: IL test D-dimer. Lab Hematol 2000;6:147 — 50. Bozˇicˇ M, Stegnar M. Validation of an automated immunoturbidimetric assay for measurement of plasma D-dimer. Clin Chem Lab Med 2003;41:958 — 62.