ORIGINAL CONTRIBUTION deep venous thrombosis; ultrasound; venography
Real-Time Ultrasound for the Detection of Deep Venous Thrombosis Purpose: Accurate diagnosis of deep venous thrombosis (DVT) is a clinical problem in emergency practice. A prospective trial was conducted comparing reaLtime ultrasound with contrast venography in the diagnosis of proximal DVT. Methods: Seventy patients whose clinical presentations mandated diagnostic evaluation .for DVT had real-time ultrasound of the involved leg followed by contrast venography Initial readings of ultrasound and venography were compared with each other and with final readings to assess reliaMlity Of interpretation. Results: Final ultrasound readings agreed with final venogram readings in all patients. Negative initial ultrasound readings agreed with final venogram readings in 56 of 56 patients (negative predictive value, 100%; 95% confidence interval, 94 to 100). Eighteen patients had positive initial ultrasound readings compared with 14 who had positive final venogram readings (positive predictive value, 78%; 95% confidence interval, 55 to 91). Conclusion: Negative real-time uJtrasonography reliably excludes proximal DVT. Positive ultrasound reliably diagnoses proximal DVT only in experienced hands. [Chance JF, Abbitt PL, Tegtmeyer CJ, Powers RD: Real time ultrasound ,for the detection of deep venous thrombosis. Ann Emerg Med May 1991;20:494-496.]
Joseph F Chance, MD* Patricia L Abbitt, MD1Charles J Tegtmeyer, MDt Robert D Powers, MD* Charlottesville, Virginia From the Departments of Internal Medicine* and Radiology,t University of Virginia School of Medicine, Charlottesville. Received for publication July 2, 1990. Revision received October 10, 1990. Accepted for publication November 19, 1990. Presented at the Society for Academic Emergency Medicine Annual Meeting in Minneapolis, Minnesota, May 1990. Address for reprints: Joseph F Chance, MD, Division of Emergency Medicine, Box 523-21, Charlottesville, Virginia 22908.
INTRODUCTION Accurate diagnosis of deep venous thrombosis (DVT) is a problem for the emergency physician. Clinical signs and symptoms are unreliable.< z Contrast venography, the gold standard, is invasive and expensive and exposes patients to radiation and the risk of contrast reactions. Several alternatives to venography have been investigated. Impedance plethysmography is accurate and safe in diagnosing proximal DVT, but availability and ease of performance limit its use.3, 4 Doppler uhrasonography is difficult to interpret, and its usefulness in diagnosing DVT has not been demonstrated. 5 Real-time B-mode ultrasound is widely available and has been studied extensively. 6-9 A large outpatient study showed virtually perfect agreement with results of venography in diagnosing proximal DVT using the single criterion of vein compressibility. Other ultrasonographic criteria such as echogenic venous contents and change in venous diameter with the Valsalva maneuver were unreliable. Isolated calf vein thrombosis is not reliably diagnosed by ultrasound. 5 Examples of an abnormal and a normal ultrasound examination are shown (Figures 1 and 2). In light of such encouraging reports, ultrasound shows promise as an emergency department procedure to diagnose DVT. To assess the usefulness of ultrasound for detecting DVT in ED patients, two issues must be addressed. Is ultrasound as useful in an ED patient population as it has proven in other outpatient populations studied? How reliable is the initial ultrasound interpretation that is used to guide clinical decisions? To assess the role of ultrasound in the ED diagnosis of DVT, a prospective study of ED patients with clinically suspected DVT was conducted, in which the results of real-time B-mode ultrasound were compared with
20:5 May 1991
Annals of Emergency Medicine
494/23
THROMBOSIS Chance et al
FIGURE 1. Normal ultrasound. A. Femoral arterial and venous lumens are labeled (arrows). B. Venous lumen has disappeared with compression. FIGURE 2. Abnormal ultrasound. A. Femoral arterial and venous lumens are labeled (arrows). B. Venous lumen fails to collapse with compression and becomes elongated, indicating presence of clot. those of contrast venography. MATERIALS A N D M E T H O D S Adult ED patients whose clinical presentations mandated evaluation for lower-extremity DVT were admitted into the study. The decision for study entry was made by the ED attending physician based on clinical suspicion of DVT. Each patient had real-time B-mode ultrasound of the involved extremity performed by the radiology resident or attending on call for ultrasound. The c o m m o n femoral and popliteal veins were evaluated for compressibility; calf veins were not examined. An initial ultrasound reading was made available to the clinician caring 24/495
for the patient. The patient then had contrast venography of the same extremity. This was performed by the radiology resident or attending on call for venography who was blinded to the results of the ultrasound reading. An initial venogram reading was reported, and a patient management decision was made by the clinician with knowledge of both test readings. The u l t r a s o u n d studies were rev i e w e d later by a f a c u l t y u l t r a sonographer, who produced a final reading. The venograms were reviewed by a faculty vascular radiologist, who produced a final reading. The patients were from an ED with an annual volume of 40,000 patients and were admitted to the study during a one-year period. Patients who had a history of contrast reactions or were pregnant were excluded. Sensitivity, specificity, and positive and negative predictive values were calculated for both initial and final readings. Confidence intervals were calculated using Confidential Interval Analysis software, to The study protocol was approved by the H u m a n Investigation Committee of the University of Virginia Health Sciences Center. Annals of Emergency Medicine
RESULTS Seventy patients were entered into the study; 20 (29%) had positive venograms on both initial and final readings. Fifty patients (71%) had negative venograms on both initial and final readings. In no case did the final venogram reading differ from the initial reading. Fourteen of 20 patients (70%) had proximal DVT. Six of 20 patients (30%) had isolated calf vein DVT. Eighteen patients (26%) had a positive ultrasound on initial reading, and 14 patients had a positive ultrasound on final reading. The four patients with initial readings of positive ultrasounds and final readings of negative u h r a s o u n d s had negative venograms. Fifty-two patients (74%) had a negative ultrasound on initial reading for proximal DVT. All 52 had negative uhrasounds on final reading and negative venograms for proximal DVT. Six of these patients had isolated calf vein thrombosis on venography. Therefore, there were 14 true-positive patients (proximal DVT on venography); 56 true-negative patients (no proximal DVT on venography, including patients with isolated calf 20:5 May 1991
THROMBOSIS Chance et al
TABLE 1. Proximal D V T
Initial ultrasound readings - -
No. of Patients % Sensitivity Specificity Positive predictivevalue Negative predictivevalue
14/14 56/60 14/18 56/56
100 93 78 100
95% CI 79-100 84 97 55 91 94--100
TABLE 2. Proximal D V T - Final ultrasound readings
No. of Patients % 95% CI Sensitivity
14/14
100 79 100
Specificity Positive predictive value Negative predictivevalue
56/56 14/14 56/56
100 94--100 100 79 100 100 94--100
vein thrombosis); four false-positive patients on initial ultrasound (initial ultrasound positive for p r o x i m a l DVT and v e n o g r a m n o r m a l ) ; no false-positive patients on final ultrasound; and no false-negative patients on both initial and final ultrasounds. The sensitivity, specificity, and positive and negative p r e d i c t i v e values for both initial and final r.eadings of ultrasound compared with venograms for proximal DVT are shown with their respective 95% confidence intervals (Tables 1 and 2). DISCUSSION The patient with possible DVT presents a diagnostic and therapeutic dilemma for the emergency physician. Clinical parameters are unreliable; only 29% of ED patients with suspected DVT had positive venograms in this study. This is similar to the 30% to 35% accuracy of clinical diagnoses reported in large outpatient studies, s Results of our study affirm the necessity to confirm clinical impression with an objective diagnostic test and suggest that clinical diagnoses are equally unreliable in ED and other outpatient populations. The large majority of symptomatic
20:5 May 1991
patients with proximal DVT in this study (70%) compared with those with isolated calf vein DVT {30%) is also c o n s i s t e n t with results from large outpatient studies, s Thus, the ED patient with DVT closely resembles patients seen in other outpatient venues.
All final u l t r a s o u n d readings in this study agreed with venograms in diagnosing proximal DVT and were consistent with the accuracy shown in other outpatient studies. 5 Therefore, clinical decisions based on final ultrasound readings alone appear to be justified. The low positive predictive value of the initial reading is troubling, and caution should be exercised when using a positive ultrasound alone as a basis for anticoagulation. A negative test is more reliable as a basis for withholding anticoagulation, even on initial reading. The incorrectly interpreted ultrasounds in our study were all falsepositives; the initial reading was not misleading if it was negative. There is unresolved controversy concerning the need to treat isolated calf vein DVT. Risks of pulmonary e m b o l i s m and p o s t p h l e b i t i c syndrome in patients with isolated calf vein DVT are hotly debated.11 However, the 20% to 30% risk of proxim a l p r o p a g a t i o n is well e s t a b lished.g, 4 A single negative ultras o u n d , no m a t t e r h o w r e l i a b l y interpreted, cannot be the basis for discharging a patient without anticoagulation or follow-up. Serial ultrasounds would be necessary to exclude propagation that would mandate a n t i c o a g u l a t i o n . Such serial testing has proven safe for impedance plethysmography. 4 Serial ultrasound should be similarly evaluated to determine the frequency and timing of repeat testing and to establish the safety of withholding anticoagulant therapy for patients with repeatedly normal ultrasounds. Results from our study support treatment of ED patients for proximal DVT based on a positive ultrasound performed by an experienced operator. Caution is indicated when
Annals of Emergency Medicine
making clinical decisions based on a positive ultrasound if experienced interpretation is not available. Negative ultrasound readings by both inexperienced and experienced operators agreed with results of venography. Proper use of ultrasound in minimizing the need for contrast venography has several potential advantages. Ultrasound is noninvasive. There is no radiation exposure. The risk of reaction to contrast media, which is as high as 3%, is avoided with its use. n In our hospital, ultrasound is much cheaper than venography (total cost: ultrasound, $161; venography, $521). CONCLUSION There is a need for noninvasive modalities to diagnose DVT in ED patients. A negative real-time B-mode ultrasound reliably excludes proximal DVT. A positive ultrasound is less reliable and detects proximal DVT only when performed by an experienced operator.
REFERENCES 1. Cranley JJ, Canos AJ, Sull WJ: The diagnosis of deep venous thrombosis: Fallibility of clinical symptoms and signs. Arch Surg 1976;111:34-36. 2. Haeger K: Problems of acute deep venous thrombosis: I. The interpretation of signs and symptoms. Angiology 1969;20:219-223. 3. Hull RD, Hixsh J, Carter CJ, et al: Diagnostic efficacy of impedance plethysmography for clinically suspected deep-vein thrombosis: A randomized trial. Ann h~tern Med 1985;102:21-28. 4. Huisman MV, B/iller HR, ten Cate JW, et ah Serial impedance plethysmography for suspected deep venous thrombosis in outpatients: The Amsterdam General Practitioner Study. N Engl J Med 1986;314:823-828. 5. Lensing AWA, Prandoni P, B~andjes D: Detection of deep-vein thrombosis by real-time B-mode ultrasonography. N Engl I Med 1989;320:342 345. 6. Sullivan ED, Peter DJ, Cranley JJ: Real-time B~mode venous ultrasound. J Vasc Surg 1984;1:465 471. 7. Langsfeld M, Hershey FB, Thorpe L, et ah Duplex B-mode imaging for the diagnosis of deep venous throm bosis. Arch Surg 1987~122:587-591. 8. Appelman PT, De Jong TE, Lampmann LE: Deep venous thrombosis of the leg: US findings. Radiology 1987;163:743-746. 9. Cronan JJ, Dorfman GS, Scola FH, et ah Deep venous thrombosis: US assessment using vein compression. Radiology 1987;162:191-i94. 10. CIA Software, Version 1.0. London, British Medical Journal, 1989. 11. Shehadi WH: Contrast media adverse reactions: Occurrence, recurrence, and distribution patterns. Radiology 1982;143:11 17.
496/25
Real-Time Ultrasound for the Detection of Deep Venous Thrombosis Purpose: Accurate diagnosis of deep venous thrombosis (DVT) is a clinical problem in emergency practice. A prospective trial was conducted comparing reaLtime ultrasound with contrast venography in the diagnosis of proximal DVT. Methods: Seventy patients whose clinical presentations mandated diagnostic evaluation .for DVT had real-time ultrasound of the involved leg followed by contrast venography Initial readings of ultrasound and venography were compared with each other and with final readings to assess reliaMlity Of interpretation. Results: Final ultrasound readings agreed with final venogram readings in all patients. Negative initial ultrasound readings agreed with final venogram readings in 56 of 56 patients (negative predictive value, 100%; 95% confidence interval, 94 to 100). Eighteen patients had positive initial ultrasound readings compared with 14 who had positive final venogram readings (positive predictive value, 78%; 95% confidence interval, 55 to 91). Conclusion: Negative real-time uJtrasonography reliably excludes proximal DVT. Positive ultrasound reliably diagnoses proximal DVT only in experienced hands. [Chance JF, Abbitt PL, Tegtmeyer CJ, Powers RD: Real time ultrasound ,for the detection of deep venous thrombosis. Ann Emerg Med May 1991;20:494-496.]
Joseph F Chance, MD* Patricia L Abbitt, MD1Charles J Tegtmeyer, MDt Robert D Powers, MD* Charlottesville, Virginia From the Departments of Internal Medicine* and Radiology,t University of Virginia School of Medicine, Charlottesville. Received for publication July 2, 1990. Revision received October 10, 1990. Accepted for publication November 19, 1990. Presented at the Society for Academic Emergency Medicine Annual Meeting in Minneapolis, Minnesota, May 1990. Address for reprints: Joseph F Chance, MD, Division of Emergency Medicine, Box 523-21, Charlottesville, Virginia 22908.
INTRODUCTION Accurate diagnosis of deep venous thrombosis (DVT) is a problem for the emergency physician. Clinical signs and symptoms are unreliable.< z Contrast venography, the gold standard, is invasive and expensive and exposes patients to radiation and the risk of contrast reactions. Several alternatives to venography have been investigated. Impedance plethysmography is accurate and safe in diagnosing proximal DVT, but availability and ease of performance limit its use.3, 4 Doppler uhrasonography is difficult to interpret, and its usefulness in diagnosing DVT has not been demonstrated. 5 Real-time B-mode ultrasound is widely available and has been studied extensively. 6-9 A large outpatient study showed virtually perfect agreement with results of venography in diagnosing proximal DVT using the single criterion of vein compressibility. Other ultrasonographic criteria such as echogenic venous contents and change in venous diameter with the Valsalva maneuver were unreliable. Isolated calf vein thrombosis is not reliably diagnosed by ultrasound. 5 Examples of an abnormal and a normal ultrasound examination are shown (Figures 1 and 2). In light of such encouraging reports, ultrasound shows promise as an emergency department procedure to diagnose DVT. To assess the usefulness of ultrasound for detecting DVT in ED patients, two issues must be addressed. Is ultrasound as useful in an ED patient population as it has proven in other outpatient populations studied? How reliable is the initial ultrasound interpretation that is used to guide clinical decisions? To assess the role of ultrasound in the ED diagnosis of DVT, a prospective study of ED patients with clinically suspected DVT was conducted, in which the results of real-time B-mode ultrasound were compared with
20:5 May 1991
Annals of Emergency Medicine
494/23
THROMBOSIS Chance et al
FIGURE 1. Normal ultrasound. A. Femoral arterial and venous lumens are labeled (arrows). B. Venous lumen has disappeared with compression. FIGURE 2. Abnormal ultrasound. A. Femoral arterial and venous lumens are labeled (arrows). B. Venous lumen fails to collapse with compression and becomes elongated, indicating presence of clot. those of contrast venography. MATERIALS A N D M E T H O D S Adult ED patients whose clinical presentations mandated evaluation for lower-extremity DVT were admitted into the study. The decision for study entry was made by the ED attending physician based on clinical suspicion of DVT. Each patient had real-time B-mode ultrasound of the involved extremity performed by the radiology resident or attending on call for ultrasound. The c o m m o n femoral and popliteal veins were evaluated for compressibility; calf veins were not examined. An initial ultrasound reading was made available to the clinician caring 24/495
for the patient. The patient then had contrast venography of the same extremity. This was performed by the radiology resident or attending on call for venography who was blinded to the results of the ultrasound reading. An initial venogram reading was reported, and a patient management decision was made by the clinician with knowledge of both test readings. The u l t r a s o u n d studies were rev i e w e d later by a f a c u l t y u l t r a sonographer, who produced a final reading. The venograms were reviewed by a faculty vascular radiologist, who produced a final reading. The patients were from an ED with an annual volume of 40,000 patients and were admitted to the study during a one-year period. Patients who had a history of contrast reactions or were pregnant were excluded. Sensitivity, specificity, and positive and negative predictive values were calculated for both initial and final readings. Confidence intervals were calculated using Confidential Interval Analysis software, to The study protocol was approved by the H u m a n Investigation Committee of the University of Virginia Health Sciences Center. Annals of Emergency Medicine
RESULTS Seventy patients were entered into the study; 20 (29%) had positive venograms on both initial and final readings. Fifty patients (71%) had negative venograms on both initial and final readings. In no case did the final venogram reading differ from the initial reading. Fourteen of 20 patients (70%) had proximal DVT. Six of 20 patients (30%) had isolated calf vein DVT. Eighteen patients (26%) had a positive ultrasound on initial reading, and 14 patients had a positive ultrasound on final reading. The four patients with initial readings of positive ultrasounds and final readings of negative u h r a s o u n d s had negative venograms. Fifty-two patients (74%) had a negative ultrasound on initial reading for proximal DVT. All 52 had negative uhrasounds on final reading and negative venograms for proximal DVT. Six of these patients had isolated calf vein thrombosis on venography. Therefore, there were 14 true-positive patients (proximal DVT on venography); 56 true-negative patients (no proximal DVT on venography, including patients with isolated calf 20:5 May 1991
THROMBOSIS Chance et al
TABLE 1. Proximal D V T
Initial ultrasound readings - -
No. of Patients % Sensitivity Specificity Positive predictivevalue Negative predictivevalue
14/14 56/60 14/18 56/56
100 93 78 100
95% CI 79-100 84 97 55 91 94--100
TABLE 2. Proximal D V T - Final ultrasound readings
No. of Patients % 95% CI Sensitivity
14/14
100 79 100
Specificity Positive predictive value Negative predictivevalue
56/56 14/14 56/56
100 94--100 100 79 100 100 94--100
vein thrombosis); four false-positive patients on initial ultrasound (initial ultrasound positive for p r o x i m a l DVT and v e n o g r a m n o r m a l ) ; no false-positive patients on final ultrasound; and no false-negative patients on both initial and final ultrasounds. The sensitivity, specificity, and positive and negative p r e d i c t i v e values for both initial and final r.eadings of ultrasound compared with venograms for proximal DVT are shown with their respective 95% confidence intervals (Tables 1 and 2). DISCUSSION The patient with possible DVT presents a diagnostic and therapeutic dilemma for the emergency physician. Clinical parameters are unreliable; only 29% of ED patients with suspected DVT had positive venograms in this study. This is similar to the 30% to 35% accuracy of clinical diagnoses reported in large outpatient studies, s Results of our study affirm the necessity to confirm clinical impression with an objective diagnostic test and suggest that clinical diagnoses are equally unreliable in ED and other outpatient populations. The large majority of symptomatic
20:5 May 1991
patients with proximal DVT in this study (70%) compared with those with isolated calf vein DVT {30%) is also c o n s i s t e n t with results from large outpatient studies, s Thus, the ED patient with DVT closely resembles patients seen in other outpatient venues.
All final u l t r a s o u n d readings in this study agreed with venograms in diagnosing proximal DVT and were consistent with the accuracy shown in other outpatient studies. 5 Therefore, clinical decisions based on final ultrasound readings alone appear to be justified. The low positive predictive value of the initial reading is troubling, and caution should be exercised when using a positive ultrasound alone as a basis for anticoagulation. A negative test is more reliable as a basis for withholding anticoagulation, even on initial reading. The incorrectly interpreted ultrasounds in our study were all falsepositives; the initial reading was not misleading if it was negative. There is unresolved controversy concerning the need to treat isolated calf vein DVT. Risks of pulmonary e m b o l i s m and p o s t p h l e b i t i c syndrome in patients with isolated calf vein DVT are hotly debated.11 However, the 20% to 30% risk of proxim a l p r o p a g a t i o n is well e s t a b lished.g, 4 A single negative ultras o u n d , no m a t t e r h o w r e l i a b l y interpreted, cannot be the basis for discharging a patient without anticoagulation or follow-up. Serial ultrasounds would be necessary to exclude propagation that would mandate a n t i c o a g u l a t i o n . Such serial testing has proven safe for impedance plethysmography. 4 Serial ultrasound should be similarly evaluated to determine the frequency and timing of repeat testing and to establish the safety of withholding anticoagulant therapy for patients with repeatedly normal ultrasounds. Results from our study support treatment of ED patients for proximal DVT based on a positive ultrasound performed by an experienced operator. Caution is indicated when
Annals of Emergency Medicine
making clinical decisions based on a positive ultrasound if experienced interpretation is not available. Negative ultrasound readings by both inexperienced and experienced operators agreed with results of venography. Proper use of ultrasound in minimizing the need for contrast venography has several potential advantages. Ultrasound is noninvasive. There is no radiation exposure. The risk of reaction to contrast media, which is as high as 3%, is avoided with its use. n In our hospital, ultrasound is much cheaper than venography (total cost: ultrasound, $161; venography, $521). CONCLUSION There is a need for noninvasive modalities to diagnose DVT in ED patients. A negative real-time B-mode ultrasound reliably excludes proximal DVT. A positive ultrasound is less reliable and detects proximal DVT only when performed by an experienced operator.
REFERENCES 1. Cranley JJ, Canos AJ, Sull WJ: The diagnosis of deep venous thrombosis: Fallibility of clinical symptoms and signs. Arch Surg 1976;111:34-36. 2. Haeger K: Problems of acute deep venous thrombosis: I. The interpretation of signs and symptoms. Angiology 1969;20:219-223. 3. Hull RD, Hixsh J, Carter CJ, et al: Diagnostic efficacy of impedance plethysmography for clinically suspected deep-vein thrombosis: A randomized trial. Ann h~tern Med 1985;102:21-28. 4. Huisman MV, B/iller HR, ten Cate JW, et ah Serial impedance plethysmography for suspected deep venous thrombosis in outpatients: The Amsterdam General Practitioner Study. N Engl J Med 1986;314:823-828. 5. Lensing AWA, Prandoni P, B~andjes D: Detection of deep-vein thrombosis by real-time B-mode ultrasonography. N Engl I Med 1989;320:342 345. 6. Sullivan ED, Peter DJ, Cranley JJ: Real-time B~mode venous ultrasound. J Vasc Surg 1984;1:465 471. 7. Langsfeld M, Hershey FB, Thorpe L, et ah Duplex B-mode imaging for the diagnosis of deep venous throm bosis. Arch Surg 1987~122:587-591. 8. Appelman PT, De Jong TE, Lampmann LE: Deep venous thrombosis of the leg: US findings. Radiology 1987;163:743-746. 9. Cronan JJ, Dorfman GS, Scola FH, et ah Deep venous thrombosis: US assessment using vein compression. Radiology 1987;162:191-i94. 10. CIA Software, Version 1.0. London, British Medical Journal, 1989. 11. Shehadi WH: Contrast media adverse reactions: Occurrence, recurrence, and distribution patterns. Radiology 1982;143:11 17.
496/25
