D-dimer is not useful for the diagnosis of isolated superficial venous thrombosis

Abstract

Combined with clinical probability assessment, D-dimer has consistently proved to be a useful diagnostic assay for the exclusion of deep venous thrombosis,1Perrier A. Desmarais S. Miron M.-J. et al.Non-invasive diagnosis of venous thromboembolism in outpatients.Lancet. 1999; 353: 190-195Abstract Full Text Full Text PDF PubMed Scopus (764) Google Scholar, 2Kearon C. Ginsberg J.S. Douketis J. et al.Management of suspected deep venous thrombosis in outpatients by using clinical assessment and D-dimer testing.Ann Intern Med. 2001; 135: 108-111Crossref PubMed Scopus (193) Google Scholar, 3Wells P.S. Anderson D.R. Rodger M. et al.Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis.N Engl J Med. 2003; 349: 1227-1235Crossref PubMed Scopus (1074) Google Scholar especially in outpatients showing lower clinical pretest probability scores.3Wells P.S. Anderson D.R. Rodger M. et al.Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis.N Engl J Med. 2003; 349: 1227-1235Crossref PubMed Scopus (1074) Google Scholar There are very few reports specifically addressing the issue of the diagnostic yield of D-dimer in patients with superficial venous thrombosis of the lower limbs, and to our knowledge none of them include an assessment of quantitative D-dimer assays in this clinical setting. In the current study we evaluated a quantitative immunoturbidimetric D-dimer assay for the diagnosis of episodes of superficial venous thrombosis diagnosed in a series of outpatients who presented with a clinical suspicion of deep venous thrombosis of the lower limbs. We carried out an observational prospective study aimed to assess the diagnostic value of D-dimer for the diagnosis of superficial venous thrombosis. Patients enrolled were individuals aged more than 18 years who presented at the emergency department of the Santa Barbara General Hospital (Soria, Spain) between November 2003 and September 2004 showing clinical symptoms suggesting a possible diagnosis of deep venous thrombosis. Exclusion criteria were as follows: diagnosis of deep venous thrombosis at presentation (either isolated or associated to superficial venous thrombosis), ongoing oral anticoagulant treatment, and pregnancy. Patients in whom a diagnosis of superficial venous thrombosis of the lower limbs was made constituted our primary target population. Superficial venous thrombosis was diagnosed in patients who presented with evidence of pain, swelling, and/or inflammation of the leg, as well as palpable distension and cord-like induration of superficial veins of the lower extremities and in whom deep venous thrombosis was ruled out. An initial clinical assessment by the attending physician was performed using the modified Wells score for suspected deep venous thrombosis.3Wells P.S. Anderson D.R. Rodger M. et al.Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis.N Engl J Med. 2003; 349: 1227-1235Crossref PubMed Scopus (1074) Google Scholar Patients scored as “deep venous thrombosis likely” or “deep venous thrombosis unlikely” were managed according to the diagnostic algorithms showed in Figure 1, Figure 2, respectively.Figure 2Diagnostic algorithm of patients classified as likely for the existence of deep venous thrombosis (DVT).View Large Image Figure ViewerDownload (PPT) We used an immunoturbidimetric D-dimer assay (STA Liatest D-Di; Diagnostica Stago, Asniéres sur Seine, France) providing quantitative results within 8 minutes. An STA-Compact analyzer (Diagnostica Stago) was used for sample testing. Samples were assayed within 30 minutes from admission. Any plasma concentration of D-dimer less than 0.4 μg/mL was regarded as a negative result as previously validated for deep venous thrombosis.4van der Graaf F. van den Borne H. van den Kolk M. et al.Exclusion of deep venous thrombosis with D-dimer testing. Comparison of 13 D-dimer methods in 99 outpatients suspected of deep venous thrombosis using venography as reference standard.Thromb Haemost. 2000; 83: 191-198PubMed Google Scholar, 5Aguilar C. Martínez A. Martínez A. et al.Diagnostic value of D-dimer in patients with a moderate pretest probability of deep venous thrombosis.Br J Haematol. 2002; 118: 275-277Crossref PubMed Scopus (22) Google Scholar In our experience and that reported by other authors, higher cutoff values do not translate into a better diagnostic yield, because they lead to a decrease in sensitivity; also, lower values reduce specificity and translate into a higher need of confirmation imaging procedures.4van der Graaf F. van den Borne H. van den Kolk M. et al.Exclusion of deep venous thrombosis with D-dimer testing. Comparison of 13 D-dimer methods in 99 outpatients suspected of deep venous thrombosis using venography as reference standard.Thromb Haemost. 2000; 83: 191-198PubMed Google Scholar, 6Aguilar C. Martínez A. Martínez A. et al.Diagnosis of deep venous thrombosis in the elderly a higher D-dimer cut-off value is better?.Haematologica. 2002; 86: E28Google Scholar In patients in whom a thrombophilia screen was considered, the complete workup included: thrombin time, functional antithrombin level, clotting protein C and protein S assays, free protein S, activated protein C resistance, FVIII levels, lupus anticoagulant, anticardiolipin and anti-beta2 glycoprotein-I antibodies, factor V Leiden, G20210A prothrombin gene mutation, C677T MTHFR gene mutation, and fasting total plasma homocysteine levels. Lower-limb Doppler venous compression ultrasonography was performed within 3 hours of admission to the emergency department by a senior radiologist as indicated according to the diagnostic algorithm (Figure 1, Figure 2) in both patients considered unlikely for deep venous thrombosis and a positive D-dimer result and all individuals classified as likely for deep venous thrombosis.3Wells P.S. Anderson D.R. Rodger M. et al.Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis.N Engl J Med. 2003; 349: 1227-1235Crossref PubMed Scopus (1074) Google Scholar, 7Keeling D.M. Mackie I.J. Moody A. Watson H.G. Haemostasis and Thrombosis Task Force of the British Committee for Standards in HaematologyThe diagnosis of deep vein thrombosis in symptomatic outpatients and the potential for clinical assessment and D-dimer assays to reduce the need for diagnostic imaging.Br J Haematol. 2004; 124: 15-25Crossref PubMed Scopus (77) Google Scholar Ultrasonographic diagnostic criteria for deep venous thrombosis have been reported,8Lensing A.W.A. Prandoni P. Brandjes D. et al.Detection of deep-vein thrombosis by real-time B-mode ultrasonography.N Engl J Med. 1989; 320: 342-345Crossref PubMed Scopus (799) Google Scholar and evidence of superficial venous thrombosis was shown in patients who underwent venous ultrasound. Both physicians examining the patients and radiologists performed an independent and blinded evaluation of the different results for each patient (D-dimer value, clinical classification, and venous ultrasound). Laboratory technicians were also unaware of the results other than D-dimer. Patients with an initial lower clinical pretest probability (deep venous thrombosis unlikely) and a negative D-dimer result were given instructions to see their general practitioners for referral to the hospital if symptoms persisted or were aggravated, and they were contacted every month for an overall period of 3 months to record any episode of deep venous thrombosis that might have been initially missed. Medical records were also reviewed at the end of that period to ascertain any new episodes of deep venous thrombosis, intercurrent disorders, or thrombotic risk factors. Patients initially classified as “deep venous thrombosis likely” and not showing any initial evidence of deep venous thrombosis had a second venous ultrasound performed within 1 week from presentation (Figure 2). All statistical parameters included in our study and their 95% confidence intervals (CIs) were calculated according to binomial distribution using Epidat (version 2.1; Xunta de Galicia, Galicia, Spain) and Statgraphic Plus (version 2.1; STSC Inc., Rockville, Md) software. We enrolled 401 outpatients presenting with clinically suspected deep venous thrombosis, of whom only 282 were eligible for evaluation; the rest of the individuals (119 patients) were excluded because of confirmed deep venous thrombosis (90 cases), ongoing oral anticoagulant treatment (25 cases), or pregnancy (4 cases). The mean age of the entire eligible population was 68 years (standard deviation 15.2), and 63.2% (178/282) of them were women. There were 35 isolated superficial venous thrombosis cases diagnosed (prevalence 12.4%). The mean age of this subset of patients was 69 years (standard deviation 15.5); women accounted for 60% (21/35) of them. More than 70% of the patients (25, 71.4%) showed at least one risk factor for thrombosis, with age more than 65 years (23 cases, 65.7%), varicose veins (15 cases, 42.8%), previous deep venous thrombosis (4 cases, 11.4%), and thrombophilia (3 cases, 8.6%) being the most commonly reported. The majority of patients with superficial venous thrombosis (32, 91.4%) were categorized as unlikely to have deep venous thrombosis, and 26 of them (74.2%) had venous ultrasonography performed as a result of the application of our diagnostic algorithm; this procedure confirmed the clinical impression in such patients. Thrombotic episodes involving the greater saphenous vein accounted for 37.1% (13/35) of patients with superficial venous thrombosis, and all but 2 of them (84.6%) showed abnormal D-dimer concentrations; all 13 individuals were considered as unlikely for the diagnosis of deep venous thrombosis. In 1 of these patients (7.7%) a progression to deep venous thrombosis was diagnosed during the follow-up period; this was the only thrombotic event recorded during this period in the entire series of patients with superficial venous thrombosis (including 3 cases who had a second ultrasound performed 1 week after the initial presentation). Sensitivity of D-dimer for the diagnosis of superficial venous thrombosis was 74.3% (95% CI 57.9%-85.8%), specificity was 43.3% (95% CI 37.3%-49.6%), positive predictive value was 15.7% (CI 10.9%-22%), and negative predictive value was 92.2% (95% CI 85.9%-95.9%). Likelihood ratios for a positive and negative D-dimer result were 1.31 (95% CI 1.05-1.64%) and 0.59 (95% CI 0.33-1.07), respectively. Post-test probability after a positive and a negative result was 15.3% (95% CI 10.6%-21.6%) and 7.6% (95% CI 4%-13.9%), respectively. The role that superficial venous thrombosis has played in the international literature has been largely eclipsed by the relevance given the thrombotic disorders of deep venous system of the lower limbs because of their frequency and their potential clinical and therapeutic implications. In our experience, deep venous thrombosis is actually more often diagnosed than isolated superficial venous thrombosis in patients with clinically suspected deep venous thrombosis (30.2% vs 12.4%) as opposed to the results reported by other authors.9Siragusa S. Terulla V. Pirelli S. et al.A rapid D-dimer assay in patients presenting at an emergency room with suspected acute venous thrombosis accuracy and relation to clinical variables.Haematologica. 2001; 86: 856-861PubMed Google Scholar Superficial venous thrombosis is more often regarded as a benign condition, but in cases involving the greater saphenous vein it can be complicated by progression of the thrombus to the deep venous system (∼8% of cases) or development of pulmonary embolism (∼10%)10Blumemberg R.M. Barton E. Gelfand M.L. et al.Occult deep vein thrombosis complicating superficial thrombophlebitis.J Vasc Surg. 1998; 27: 338-343Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar; consistent with these figures, approximately 8% of our patients with greater saphenous vein superficial venous thrombosis developed deep venous thrombosis during the follow-up period, but none presented with symptoms consistent with pulmonary embolism. These thrombotic episodes involving the superficial venous system can be associated with risk factors such as varicose veins, obesity, thrombophilia, pregnancy, malignant disorders, or chronic myeloproliferative syndromes.11Unno M. Mitsuoka H. Uchiyama T. et al.Superficial thrombophlebitis of the lower limbs in patients with varicose veins.Surg Today. 2002; 32: 397-401Crossref PubMed Scopus (41) Google Scholar, 12de Moerloose P. Wutschert R. Heinzmann M. et al.Superficial vein thrombosis of the lower limbs influence of factor V Leiden, factor G20210A and overweight.Thromb Haemost. 1998; 80: 239-241PubMed Google Scholar, 13McColl M.D. Ramsay J.E. Tait R.C. et al.Superficial vein thrombosis incidence in association with pregnancy and prevalence of thrombophilic defects.Thromb Haemost. 1998; 79: 741-742PubMed Google Scholar, 14Martinelli I. Mannucci P.M. De Stefano V. et al.Different risks of thrombosis in four coagulation defects associated with inherited thrombophilia a study of 150 families.Blood. 1998; 92: 2353-2358PubMed Google Scholar As we mentioned, published research on superficial venous thrombosis has been very limited for the time being, this fact becomes even more evident for the diagnostic aspects of the condition. As opposed to deep venous thrombosis, data on studies that focus on the diagnostic role of D-dimer in patients with superficial venous thrombosis are very scarce and based on nonquantitative (semiquantitative or qualitative) D-dimer assays. The study by Siragusa et al9Siragusa S. Terulla V. Pirelli S. et al.A rapid D-dimer assay in patients presenting at an emergency room with suspected acute venous thrombosis accuracy and relation to clinical variables.Haematologica. 2001; 86: 856-861PubMed Google Scholar is partly similar in design to ours, but they do not include a clinical assessment of the patients according to any established scores (because all the patients underwent venous compression ultrasound). They use a semiquantitative D-dimer assay, they only consider superficial venous thrombosis episodes arising in the most proximal segment of the saphenous system, and they do not clearly exclude patients with coexistent deep venous thrombosis; however, they provide the largest series of patients with superficial venous thrombosis available. The number of patients enrolled in the present study is limited but clearly representative of cases with isolated superficial venous thrombosis. Although not all our patients underwent compression venous ultrasound, avoidance of any diagnostic procedures in individuals with a low clinical probability and a negative D-dimer result has consistently proven to be virtually 100% safe.1Perrier A. Desmarais S. Miron M.-J. et al.Non-invasive diagnosis of venous thromboembolism in outpatients.Lancet. 1999; 353: 190-195Abstract Full Text Full Text PDF PubMed Scopus (764) Google Scholar, 2Kearon C. Ginsberg J.S. Douketis J. et al.Management of suspected deep venous thrombosis in outpatients by using clinical assessment and D-dimer testing.Ann Intern Med. 2001; 135: 108-111Crossref PubMed Scopus (193) Google Scholar, 3Wells P.S. Anderson D.R. Rodger M. et al.Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis.N Engl J Med. 2003; 349: 1227-1235Crossref PubMed Scopus (1074) Google Scholar, 5Aguilar C. Martínez A. Martínez A. et al.Diagnostic value of D-dimer in patients with a moderate pretest probability of deep venous thrombosis.Br J Haematol. 2002; 118: 275-277Crossref PubMed Scopus (22) Google Scholar, 7Keeling D.M. Mackie I.J. Moody A. Watson H.G. Haemostasis and Thrombosis Task Force of the British Committee for Standards in HaematologyThe diagnosis of deep vein thrombosis in symptomatic outpatients and the potential for clinical assessment and D-dimer assays to reduce the need for diagnostic imaging.Br J Haematol. 2004; 124: 15-25Crossref PubMed Scopus (77) Google Scholar Therefore, this diagnostic approach may be considered as a surrogate of venous ultrasonography in this particular subset of patients. We found that the majority of confirmed cases of superficial venous thrombosis were scored as unlikely for the existence of deep venous thrombosis (91.4%) by their attending physicians and that 3 of every 4 of them had elevated D-dimer levels and, therefore, underwent compression venous ultrasound. This elevation could be the result of associated conditions as well as the superficial venous thrombosis episode itself. In individuals regarded as unlikely for the existence of deep venous thrombosis at presentation and in whom this diagnostic procedure was omitted because of normal D-dimer levels, the diagnosis was supported by clinical findings and exclusion of deep venous thrombosis based on a strongly validated diagnostic approach including clinical assessment, D-dimer result, and follow-up.1Perrier A. Desmarais S. Miron M.-J. et al.Non-invasive diagnosis of venous thromboembolism in outpatients.Lancet. 1999; 353: 190-195Abstract Full Text Full Text PDF PubMed Scopus (764) Google Scholar, 3Wells P.S. Anderson D.R. Rodger M. et al.Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis.N Engl J Med. 2003; 349: 1227-1235Crossref PubMed Scopus (1074) Google Scholar, 5Aguilar C. Martínez A. Martínez A. et al.Diagnostic value of D-dimer in patients with a moderate pretest probability of deep venous thrombosis.Br J Haematol. 2002; 118: 275-277Crossref PubMed Scopus (22) Google Scholar Diagnostic procedures have been considered unnecessary for confirmation of superficial venous thrombosis, but the accuracy of clinical diagnosis has not been validated yet and coexistence of deep venous thrombosis has been reported in 7% of patients in whom superficial venous thrombosis affects the greater saphenous territory. For these reasons, we share the view of other authors that diagnostic strategies similar to those used for deep venous thrombosis should be used in all patients in whom superficial venous thrombosis is suspected to confirm the diagnosis and exclude the not negligible proportion of patients who may show extension to the deeper venous system.9Siragusa S. Terulla V. Pirelli S. et al.A rapid D-dimer assay in patients presenting at an emergency room with suspected acute venous thrombosis accuracy and relation to clinical variables.Haematologica. 2001; 86: 856-861PubMed Google Scholar, 15Siragusa S. Quartero L. Should superficial vein thrombosis of the proximal greater saphenous vein be objectively evaluated in emergency wards?.Thromb Haemost. 2000; 83: 962-963PubMed Google Scholar A negative predictive value of D-dimer less than 98% has been considered suboptimal for safely excluding deep venous thrombosis.4van der Graaf F. van den Borne H. van den Kolk M. et al.Exclusion of deep venous thrombosis with D-dimer testing. Comparison of 13 D-dimer methods in 99 outpatients suspected of deep venous thrombosis using venography as reference standard.Thromb Haemost. 2000; 83: 191-198PubMed Google Scholar As opposed to deep venous thrombosis, the sensitivity and negative predictive value of our immunoturbidimetric D-dimer assay were far from ideal for exclusion of superficial venous thrombosis; the lower prevalence of superficial venous thrombosis in our limited series of patients might have raised the negative predictive value of D-dimer elicited. Specificity and positive predictive value were too low to be useful for confirmation, and likelihood ratios elicited were unable to change pretest probability in a clinically relevant way. Some reports have found a correlation between D-dimer levels and the extension of the thrombus16Freyburger G. Trillaud H. Labrouche S. et al.D-dimer strategy in thrombosis exclusion. A gold standard study in 100 patients suspected of deep venous thrombosis or pulmonary embolism: 8 methods compared.Thromb Haemost. 1998; 79: 32-37PubMed Google Scholar; consequently the high rate of false negative results observed (∼25%) might be explained by the low thrombotic burden associated with superficial venous thrombosis.15Siragusa S. Quartero L. Should superficial vein thrombosis of the proximal greater saphenous vein be objectively evaluated in emergency wards?.Thromb Haemost. 2000; 83: 962-963PubMed Google Scholar The proportion of false-negative results reported by other authors9Siragusa S. Terulla V. Pirelli S. et al.A rapid D-dimer assay in patients presenting at an emergency room with suspected acute venous thrombosis accuracy and relation to clinical variables.Haematologica. 2001; 86: 856-861PubMed Google Scholar, 17Jacq F. Heron E. Rance A. et al.Evaluation of a test for rapid detection of D-dimers for the exclusion of the diagnosis of venous thrombosis.Presse Med. 1997; 26: 1132-1134PubMed Google Scholar is much higher (up to 52%), but we share their conclusion that D-dimer does not significantly improve or simplify the diagnostic strategies of superficial venous thrombosis. The type of D-dimer assay used may partly account for these differences, because none of the other authors have used quantitative tests before. The majority of superficial venous thrombosis cases affecting the greater saphenous vein show abnormal D-dimer levels, which along with the high rate of false-negative results found for this parameter in this setting greatly reduce the value of a normal D-dimer concentration for exclusion of this condition. Therefore, most cases of isolated superficial venous thrombosis are typically diagnosed in patients who present with a low pretest probability of deep venous thrombosis (deep venous thrombosis unlikely) in whom deep venous thrombosis must be ruled out by a compression venous ultrasound only if D-dimer levels are above normal. Associated conditions are not uncommon and may be responsible for the elevation of D-dimer plasma concentration. D-dimer is not a useful tool for superficial venous thrombosis exclusion, and clinical criteria play a key role in this setting.