Deep Vein Thrombosis: Can a Second Sonographic Examination Be Avoided?

Alfonsa Friera 1 Nuria R. Giménez 2 Paloma Caballero 1 Pilar S. Moliní 2 Carmen Suárez 2 Received August 15, 2001; accepted after revision October 16, 2001. 1 Radiology Department, Hospital de la Princesa, Diego de León, 62 28006 Madrid, Spain. Address correspondence to A. Friera. 2 Internal Medicine Department, Hospital de la Princesa, 62 28006 Madrid, Spain. AJR 2002;178:1001 1005 0361 803X/02/1784 1001 American Roentgen Ray Society Deep Vein Thrombosis: Can a Second Sonographic Examination Be Avoided? OBJECTIVE. Our objective was to determine the prevalence of deep vein thrombosis in symptomatic patients and its distribution based on the assessment of prior clinical probability. We evaluated whether repeated sonography is necessary in patients with either intermediate or high clinical probability for deep vein thrombosis after an initial examination with negative findings. SUBJECTS AND METHODS. We prospectively evaluated 438 consecutive patients with clinical suggestion of deep vein thrombosis of the lower limbs, classified according to the prior clinical probability (high, intermediate, low). Sonography with positive findings was diagnostic for deep vein thrombosis. Negative findings in low-risk patients excluded thrombosis. Patients with intermediate or high clinical risk whose initial sonographic examination showed negative findings underwent a second examination after 1 week. RESULTS. Of the 438 patients with clinical symptoms, 112 patients (26%) had positive findings on sonography, and 326 (74%) had negative findings. Of the 202 intermediate- and high-risk patients with negative initial sonography, 140 patients underwent a single follow-up sonographic examination 1 week later. In three cases, findings were positive for deep vein thrombosis. Two other patients developed pulmonary embolism. Sonographic follow-up increased the detection of deep vein thrombosis in the patients with intermediate or high probability from 32.5% to 33.5%; the prevalence of thromboembolic disease in this group was 34%. CONCLUSION. The prevalence of deep vein thrombosis studied by sonography in the patients with intermediate or high clinical risk was 33.5%. Initial sonography revealed a 32.5% prevalence, and a second examination 1 week later detected an additional 1%. Sonography did not reveal 0.5% of thromboembolic events. Our results do not justify a routine second scanning at 1 week. T he exact incidence of deep vein thrombosis in the general population is not well known, but was estimated to be 1.6 of every 1,000 inhabitants per year in a study performed in Sweden in 1992 [1]. Pulmonary embolism constitutes the cause of death in 10% of autopsies. Deep vein thrombosis has been found at autopsy in 83% of patients with pulmonary embolism, although only 19% of them presented symptoms before death and findings were positive in only 3% of diagnostic tests that were done for deep vein thrombosis before autopsy [2]. These figures reveal not only the difficulty of diagnosing deep vein thrombosis and pulmonary embolism, but also the high incidence of asymptomatic and undiagnosed patients, in whom the assessment of the clinical probability is basic to arriving at the diagnosis [2]. Venography has been considered the reference standard for the diagnosis of deep vein thrombosis. However, this method presents several disadvantages: It requires an equipped radiology facility, it is an invasive test, it has potential side effects, and an adequate venogram cannot be obtained in 10 20% of patients for technical reasons. In the past few years, sonography has become the method of choice for the investigation of deep vein thrombosis, and some recent series have considered it to be the reference standard [3, 4]. Comparative studies with venography showed that sonography had a sensitivity of 96% and a specificity of 98% for the detection of thrombi in the femoral and popliteal systems [1, 5]. However, diagnostic accuracy of sonography for the distal calf vein thrombosis is 88 95%, and calf vein thrombosis occurs in 13 15% of deep vein thrombosis cases [6, 7]. A proportion of the isolated calf clots may extend into the proximal veins during a period of 1 week [2, 7]. Other researchers have proposed diagnostic strategies that combine the assessment of prior clinical probability, sonography, and the deter- AJR:178, April 2002 1001

Friera et al. mination of D-dimer, a fragment specific to the degradation of fibrin [8], with varying results. When the first examination is negative for deep vein thrombosis, some investigators recommend a repeated sonographic study between 1 and 14 days later [7]. This practice is expensive and causes inconvenience to the patient. For this reason, other authors suggest that repetition of sonography could be avoided. Birdwell et al. [7] studied 342 patients with clinical suggestion of deep vein thrombosis. Those with initial negative sonography findings underwent follow-up scanning; only 2% of these findings were positive. Clinical follow-up in patients with normal findings did not reveal pulmonary embolism events, and the prevalence of thromboembolic disease was 0.6%. Birdwell et al. proposed a single repeated sonographic examination 5 7 days after an initial negative study. Wells et al. [9] repeated a sonographic examination in patients who had intermediate or high clinical risk for deep vein thrombosis and an initial normal study and found a 1.2% risk for thromboembolic disease. Serum D-dimer values have proved to be sensitive and to have a high negative predictive value for the detection of deep vein thrombosis [3]. Bernardi et al. [10] established that, in patients with initial negative findings on sonography and normal D-dimer values, further sonography can be safely excluded, because they found during follow-up that venous complications occurred in only 0.2% of patients. Thromboembolic disease appeared in 9% of patients when a positive D- dimer determination appeared. The purpose of our study was, first, to determine the prevalence of deep vein thrombosis detected by sonography of the lower limbs in patients with clinical suggestion of deep vein thrombosis and to determine the distribution related to prior clinical probability. Our second objective was to define the number of positive second sonographic examinations performed after 1 week, or the prevalence of thromboembolic events in patients with intermediate or high risk and initially negative sonography. We also evaluated whether routine repeated sonography at 7 days is justified. Subjects and Methods During a 9-month period, from May 2000 through February 2001, we prospectively studied 438 consecutive patients admitted to our institution because they had clinical suggestion of deep vein thrombosis. Of these patients, 404 were evaluated in the emergency department, 30 were inpatients of various departments, and four other patients presented at the outpatient clinic. Abnormal n = 112 Anticoagulant therapy initiates Clinical suspicion of DVT n = 438 Prior clinical probability n = 438 Sonography Low n = 124 Rules out DVT Normal n = 137 Rules out DVT Patients underwent clinical assessment before any other test. Clinical probability of deep vein thrombosis was determined and subjects were classified as low-, intermediate-, or high-risk patients on the basis of risk factors for deep vein thrombosis, symptoms and signs, and the likelihood of an alternative diagnosis, in accordance with Wells et al. [11]. All patients underwent sonography of the lower limbs. Each examination was performed on a Corevision (Toshiba; Otawara-Shi, Japan) sonography unit and a 7.5-MHz linear transducer. The femoral venous system was explored with the patient in the supine position with external rotation of the leg and a slightly flexed knee. The popliteal venous system was studied with the patient in the prone or lateral decubitus position with a slight flexion of the knee. The transducer was passed transversely from the groin to the popliteal fossa with gentle pressure over the vein to cause the lumen to collapse in the absence of thrombus [12]. Color Doppler sonography helped reveal venous vessels, but it was not mandatory. Sonography was performed by a general radiologist or a third- or fourth-year radiology resident. The criteria for the presence of deep vein thrombosis included inability to compress the lumen of the vein fully with the sonography transducer or the presence of an intraluminal filling defect inside the vein. Thrombosis was excluded if both the femoral and popliteal veins were fully compressible and no residual lumen was seen. Sonographic findings of deep vein thrombosis established the diagnosis of venous thrombosis and warranted anticoagulant treatment. Patients with normal sonographic findings were treated separately, according to the clinical probability. Those patients with low clinical risk and negative examination findings for deep vein thrombosis were excluded. Patients with intermediate or high clinical probability of deep vein thrombosis and whose initial sonography findings were negative and suggested no alternative diagnosis underwent a second sonographic examination by an experienced vascular radiologist after 7 days (Fig. 1). If a patient refused the examination, we followed up clinically for 3 months. Normal n = 326 Assessment of clinical probability Intermediate or high n = 202 Repeat sonography in 7 days n = 140 Abnormal n = 3 Anticoagulant therapy initiates Fig. 1. Flowchart shows method of diagnosis of deep vein thrombosis at our institution. DVT= deep vein thrombosis. 1002 AJR:178, April 2002

Sonography of Deep Vein Thrombosis Fig. 2. 83-year-old woman with high clinical risk of deep vein thrombosis. A and B, Transverse gray-scale sonograms without compression (A) and with compression (B) reveal compressible left popliteal vein (arrow, B). C, Transverse gray-scale sonogram with compression at same level as A but obtained 1 week later shows intraluminal filling defect inside left popliteal vein (arrow). Results Of the 438 patients with clinical suggestion of deep vein thrombosis and prior clinical probability, 154 (35%) were male and 284 (65%) were female; their ages ranged from 9 to 100 years (median age, 66 years). Demographic data for each of the three risk groups were similar to that of the group as a whole. The results of initial sonography were normal in 326 patients (74%) and abnormal in 112 patients (26%). Overall prevalence of deep vein thrombosis was 26% in the 438 patients. Low Clinical Risk Patients Of the 138 patients (31.5%) with low clinical risk, 67% were female and 33% were male; their mean age was 63 years. Sonography detected deep vein thrombosis in 14 patients (10%), and in 124 patients (90%), the examination was normal. Alternative sonographic diagnoses were present in 13 patients without deep vein thrombosis: six Baker s cysts, six superficial vein thrombosis or thrombophlebitis, and one cellulitis. No routine follow-up sonography was performed in this group. Four patients were readmitted for deep vein thrombosis symptoms: in two patients, sonographic findings were normal; venous thrombosis was revealed in the other two patients (1.6%). The total prevalence of deep vein thrombosis in the low clinical risk group was 11.7%. A Intermediate Clinical Risk Patients Of the 142 patients (32.5%) with intermediate clinical risk, 65% were female and 35% were male; their mean age was 66 years. Sonography was positive for deep vein thrombosis in 30 patients (21%), and in 112 patients (79%) it was normal. Alternative sonographic diagnoses were detected in 19 patients with no deep vein thrombosis: seven Baker s cysts, three thrombophlebitis, six cellulitis, one pyomyositis, one angioedema, and one tumor. We considered the initial sonographic examination diagnostic in 49 patients (34.5%) in this group (30 patients with deep vein thrombosis and 19 patients with alternative diagnoses) (Table 1). Of the 93 patients whose findings were normal on the initial sonographic test, 72 (77%) underwent a repeated examination in 7 days: 71 had normal results, but in one patient (1.4%) TABLE 1 Risk Note. DVT = deep vein thrombosis. B sonography revealed deep vein thrombosis. Clinical information was available for 14 of the 21 patients who did not come for a second examination, and no thromboembolic events occurred in those patients. One patient, who was on anticoagulant therapy, died 6 days after the initial sonography; the probable cause of death was considered to be either bronchoaspiration or bronchospasm. Six other patients were missing for follow-up. The total prevalence of deep vein thrombosis in the intermediate risk group was 22%, and a repeated sonographic scan helped to detect 1% more than the initial examination (Table 2). Comparison of Sonographic Results with Assessment of Prior Clinical Probability for Deep Vein Thrombosis No. of Patients (%) First Studies Diagnostic of DVT (%) First Studies Nondiagnostic of DVT (%) Without Sonographic Alternative Diagnosis With Sonographic Alternative Diagnosis High 158 (36) 68 (43) 86 (54.5) 4 (2.5) Intermediate 142 (32.5) 30 (21) 93 (65.5) 19 (13.5) Low 138 (31.5) 14 (10) 111 (80.5) 13 (9.5) Total 438 (100) 112 (25.5) 290 (66) 36 (8.5) TABLE 2 Clinical Risk Comparison of First and Second Sonographic Studies in Patients with Intermediate and High Clinical Probability for Deep Vein Thrombosis Note. DVT = deep vein thrombosis. First Studies Diagnostic of DVT (%) First and Second Studies Diagnostic of DVT (%) Difference (%) High 68 (43) 70 (44) 2 (1) Intermediate 30 (21) 31 (22) 1 (1) Total 98 (32.5) 101 (33.5) 3 (1) C AJR:178, April 2002 1003

Friera et al. High Clinical Risk Patients Of the 158 patients (36%) with high clinical risk, 63% were females and 37% were males; their mean age was 67 years. Sonography revealed deep vein thrombosis in 68 patients (43%); findings were normal in 90 patients (57%). Four patients had alternative diagnoses: one Baker s cyst, one thrombophlebitis, one rhabdomyolysis, and one non-hodgkin s lymphoma. Therefore, initial sonography was diagnostic in 72 patients (45.5%) (Table 1). Of the 86 patients with normal findings on initial sonography, 67 (78%) underwent a repeated test after 1 week. In 65 patients (97%), the results remained negative; in two patients (3%), findings were positive for deep vein thrombosis (Fig. 2). Clinical follow-up did not reveal thromboembolic events in 14 of the 19 patients who did not return for repeated testing; pulmonary embolism occurred in one patient seven days after the initial sonography and in another patient 19 days after sonography. Three patients were missing for follow-up. The follow-up sonography at 1 week in this group detected 1% more deep vein thrombosis than did the initial examinations (Fig. 2). The overall prevalence of thromboembolic disease in this group was 45.5%. In summary, of the 300 patients with intermediate or high clinical probability for deep vein thrombosis, initial sonography revealed thrombi in 98 (32.5%). In five patients, further thromboembolic events appeared (deep vein thrombosis or pulmonary embolism). The total prevalence of deep vein thrombosis was 34%, and a second sonographic examination detected only 1% more. Discussion Several strategies have been developed for the diagnosis of deep vein thrombosis, including the assessment of prior clinical probability, serial sonography, and plasma values of D-dimer [8]. Sonography is accurate for revealing thrombi in the femoral and popliteal systems, and, in some instances, sonography has been considered the reference standard [3, 4]. However, for diagnosing thrombosis confined to the deep veins of the calf, sonography is less sensitive (40 70%) [6]. Serial testing, within a period of 1 7 days, has been proposed to identify patients who develop extension of thrombosis into the proximal veins. However, this approach is expensive, because 80% of patients whose symptoms suggest deep vein thrombosis have negative findings on the initial sonographic examination and would therefore need a repeated scan [7]. Gottlieb and Widjaja [13] found only two thromboembolic events (1.4%) among 146 patients with clinical suggestion of deep vein thrombosis and initial normal sonography. Cogo et al. [14] studied 1,702 patients suspected of having deep vein thrombosis: findings on initial compression sonography were normal in 1,302 patients, and the examination was repeated in 1 week. In 0.9% of these patients, the second examination had positive findings for deep vein thrombosis. Further clinical follow-up for 3 months revealed 0.7% of thromboembolic events. Birdwell et al. [7] found seven patients with positive findings on repeated sonography among 342 patients with initial normal findings; the clinical follow-up did not reveal further cases of pulmonary embolism, and the prevalence of thromboembolic disease was 0.6%. Birdwell et al. and Cogo et al. have recommended a single repeated sonographic examination 5 7 days after negative findings on an initial scan. Harper et al. [15] did not report thromboembolic events after an initial examination with normal findings, and they suggested that repetition should be limited to patients with persistent symptoms. Other researchers [16] found that such a low rate of abnormal findings on repeated sonography did not justify a systematic second examination. Diagnostic strategies, including clinical score and D- dimer testing, should be introduced. Wells et al. [9] have associated the assessment of prior clinical risk to sonography. Those patients with an initial negative scan and moderate clinical risk had a repeated sonography in 1 week; those with an initial negative scan and high clinical risk underwent venography. Long-term follow-up revealed 0.6% of thromboembolic events, which is less than the 1.9% of events reported for groups initially studied with venography [17]. On the basis of our study, the thromboembolic risk for patients who do not undergo a second sonographic study would be 1.5%. We selected the patients with initial negative findings on sonography who were of intermediate and high clinical probability according to the classification of Wells et al. [11]. During follow-up, five patients experienced thromboembolic episodes, and three cases of deep vein thrombosis were diagnosed on follow-up sonography 1 week later. Two patients had pulmonary embolisms; one occurred on the seventh day after the initial symptoms, before the patient would have undergone a second examination. In the other patient, the pulmonary embolism appeared on day 19, and follow-up sonography at 1 week might have detected a deep vein thrombosis. In our experience, only four patients (2.2%) of the total in our study could benefit from repeated sonography. Our data resemble those of Birdwell et al. [7]. Therefore, it may be true that repeated sonography is not the best method for following up patients with negative findings on the initial sonography. In some series [10], 50% of ultimate thromboembolic events were not diagnosed by a serial scan. Alternative diagnostic strategies should be used with such patients. If we analyze separately the intermediate- and high-risk groups, the increase of prevalence of deep vein thrombosis detected by a second examination was 1% for both. We believe that these results do not justify a routine repetition of sonography. In our institution, we do not include below-knee venous sonography as part of the routine normal practice, and perhaps the positive findings on second sonograms detected in our study are caused by the proximal extension of calf vein thrombi. The sonographic examination of the calf region increases the initial prevalence of deep vein thrombosis [14]. Cornuz et al. [6] found a 15% of prevalence of calf vein thrombosis when below-knee sonography was systematically practiced. This practice further reduces thromboembolic events, and a normal examination excludes deep vein thrombosis. Bernardi et al. [10] have proposed eliminating the second examination in patients with normal D-dimer values. In the future, D-dimer determination will probably reduce the number of initial and serial sonographic examinations. In summary, the prevalence of deep vein thrombosis in intermediate- and high-risk groups in our study that was detected by sonography was 33.5%. The prevalence after a second examination increased 1% in each group. Our results support the conclusion that routine repetition of sonography for the diagnosis of deep vein thrombosis is not justified. Sonography did not reveal all of the thromboembolic events. New diagnostic strategies should be introduced to follow these patients. Acknowledgments We thank the members of the Thromboembolic Disease Working Group at La Princesa Hospital. References 1. Lensing AWA, Prandoni P, Prins MH, Büller HR. Deep-vein thrombosis. Lancet 1999;353:479 485 2. Cronan JJ. Venous thromboembolic disease: the role of US. Radiology 1993;186:619 630 1004 AJR:178, April 2002

Sonography of Deep Vein Thrombosis 3. Bradley M, Bladon J, Barker H. D-Dimer assay for deep vein thrombosis: its role with colour Doppler sonography. Clin Radiol 2000;55:525 527 4. Duwe KM, Shiau M, Budorick NE, Austin JHM, Berkmen YM. Evaluation of the lower extremity veins in patients with suspected pulmonary embolism: a retrospective comparison of helical CT, venography and sonography. AJR 2000;175: 1525 1531 5. Perrier A, Desmarais S, Miron MJ, et al. Non-invasive diagnosis of venous thromboembolism in outpatients. Lancet 1999;353:190 195 6. Cornuz J, Pearson SD, Polak JF. Deep venous thrombosis: complete lower extremity venous US evaluation in patients without known risk factors: outcome study. Radiology 1999;211:637 641 7. Birdwell BG, Raskob GE, Whitsett TL, et al. The clinical validity of normal compression ultrasonography in outpatients suspected of having deep venous thrombosis. Ann Intern Med 1998;128:1 7 8. Bounameaux H, Perrier A. Rapid diagnosis of deep vein thrombosis in symptomatic patients: a comparison between four different diagnostic strategies. (letter) Thromb Haemost 1999;82:1360 1361 9. Wells PS, Anderson DR, Bormanis J, et al. Value of assessment of pretest probability of deep-vein thrombosis in clinical management. Lancet 1997; 350:1795 1798 10. Bernardi E, Prandoni P, Lensing AWA, et al. D- Dimer testing as an adjunct to ultrasonography in patients with clinically suspected deep vein thrombosis: prospective cohort study. BMJ 1998; 317:1037 1040 11. Wells PS, Hirsh J, Anderson DR, et al. Accuracy of clinical assessment of deep-vein thrombosis. Lancet 1995;345:1326 1329 12. Fraser JD, Anderson DR. Deep venous thrombosis: recent advances and optimal investigation with US. Radiology 1999;211:9 24 13. Gottlieb RH, Widjaja J. Clinical outcomes of untreated symptomatic patients with negative findings on sonography of the thigh for deep vein thrombosis: our experience and a review of the literature. AJR 1999;172:1601 1604 14. Cogo A, Lensing AWA, Koopman MMW, et al. Compression ultrasonography for diagnostic management of patients with clinically suspected deep vein thrombosis: prospective cohort study. BMJ 1998;316:17 20 15. Harper PL, Watson L, Bannon R. Compression ultrasonography for diagnosing deep vein thrombosis: protocol is safe. (comment) BMJ 1998;316:1534 16. Bounameaux H, Perrier A. Compression ultrasonography for diagnosing deep vein thrombosis: repeat testing is unjustified. (comment) BMJ 1998;316:1534 1535 17. Kraaijenhagen RA, Lensing AWA, Lijmer JG, et al. Diagnostic strategies for the management of patients with clinically suspected deep-vein thrombosis. Curr Opin Pulm Med 1997;3:268 274 AJR:178, April 2002 1005