Clinical Radiology (1987) 38, 309-313

Real-time Ultrasound Diagnosis of Deep Vein Thrombosis" A Comparison with Venography A. G. F. AITKEN and D. J. GODDEN*

Department of Radiology, Raigmore Hospital, Inverness and *Department of Medicine, University o f Aberdeen, Scotland

A prospective study comparing realltime ultrasound scanning with contrast venography in the diagnosis of deep venous thrombosis of the lower limb was per- formed in a group of 46 patients. The sensitivity of ultrasound scanning for thrombus within the ilio- femoral segment, femoral vein, and popliteal vein was 94% with a specificity of 100%. Additional information obtained by ultrasound included the diagnosis of popliteal cysts, pelvic and inguinal lymphadenopathy, popliteal haematoma, and traumatic arterial aneurysm. Realltime ultrasound scanning is a rapid and nonlinvasive alternative to contrast venography in the diagnosis of lower limb deep venous thrombosis.

The clinical assessment of suspected deep venous thrombosis in the lower limb is unreliable, with a tendency to over diagnose (Simpson et al., 1980; Hull et al., 1981). Venography is generally considered the most accurate method of diagnosis, though subject to problems of interpretation (Thomas, 1972; Browse, 1978) and is associated with complications (Albrecht- sson and Olsson, 1976, 1979; Hull et al., 1981; Kristiansen et al., 1981), which are however less frequent using modern contrast media (Thomas et al., 1984). A number of alternative procedures have been developed in an attempt to satisfy the demand for an accurate, non-invasive method of diagnosis. These include impedance plethysmography (Moser et al., 1977; Sandler et al., 1984), thermography (Pochac- zevsky et al., 1982), Doppler ultrasound (Flanigan et al., 1978; Schroeder and Dunn, 1982), Iodine - 125 fibrinogen scanning (Moser et al., 1977; Simpson et al., 1980) and recently real-time ultrasound scanning (Sullivan et al., 1984; Effeney et al., 1984; Raghavendra et al., 1984). We report a prospective study comparing real-time ultrasound scanning with contrast venography in patients clinically suspected of having lower limb deep vein thrombosis.

METHODS

Forty-six patients with clinically suspected deep vein thrombosis of the lower limb were referred for real- time ultrasound scanning and contrast venography. The ultrasound examinations were carried out by one operator (A.G.F.A.) using a real-time ultrasound unit (Advanced Technology Labs, Bellevue, WA) with a high resolution 7.5 MHz transducer. The ilio-femoral segment and femoral vein were visualised down to the

Address for correspondence: Dr A. G. F. Aitken, Department of Radiology, Raigmore Hospital, Inverness IV2 3UJ.

adductor canal region in the supine position. The popliteal vein was examined with the patient in a prone position. If visualisation of the veins was felt to be suboptimal, then the patient was also scanned while in the erect position. Sagittal and transverse images were obtained. Criteria used to determine the pre- sence of thrombosis included size and compressibility of the vein, and an assessment of intraluminal echogenicity. To aid in the assessment of lumen echogenicity the time-gain compensation curve was adjusted so that the lumen of the artery adjacent to the vein was echo free. This was done to minimise artefact echoes. Contrast venography was performed using a standard technique from a venepuncture on the dorsum of the foot and was carried out on the same day as the ultrasound examination. The radio- graphic criterion used for a positive diagnosis of venous thrombosis was the presence of an intra- luminal filling defect which was constant on all films (Thomas, 1972; Williams, 1973). Each venogram was evaluated for the presence of thrombus in the ilio- femoral segment, and the femoral and popliteal veins by one of three radiologists.

RESULTS

Venography was unsuccessful due to failure to cannulate a vein in three cases, and one ultrasound examination was inconclusive. Direct comparison of ultrasound scans with venograms was therefore pos- sible in 42 cases. Sixteen venograms were positive and 26 were negative. Fifteen ultrasound scans were posi- tive and 27 negative. Sensitivity of ultrasound was 94%, specificity 100% and predictive accuracy 100%. Additional information was obtained by ultrasound scanning in seven cases (Table 1).

Table 1 - ,Addit ional information obtained by ultrasound from seven patients

Diagnos& Namber of Cases

Popliteal cyst 3 Pelvic Lymphadenopathy 2 Haematoma 1 Traumatic arterial aneurysm 1

DISCUSSION

The normal femoral and popliteal veins appear flee of intraluminal soft tissue echoes and are easily compressed when examined by a high resolution real-time ultrasonic transducer (Fig. 1). When throm-

310 CLINICAL RADIOLOGY

(a) (a)

(b)

Fig. I - Longitudinal scan of a normal femoral vein (arrows) lying posterior to the femoral artery (a) without compression (b) with compression. The lumen can be completely closed,

(b)

bus is present within the vein a relatively echogenic intraluminal soft tissue mass is seen which prevents the vein from collapsing with compression (Fig. 2). The vein may also appear swollen if completely occluded. The normal response of an increase in the diameter of the femoral vein with the Valsalva manoeuvre is diminshed or absent with the completely occluded vein (Effeney et al., 1984). This latter feature may be a useful adjunct although we have not found this sign to add to our diagnostic accuracy. In all our cases the popliteal and ilio-femoral segments were easily de- monstrated and the effects of compression noted.

Visualisation of the femoral vein and assessment of the effects of compression was possible in all cases, although in obese individuals and in the presence of gross oedema, visualisation was improved if the pa- tient was scanned in the erect position. Visualisation of the popliteal-femoral segment in the region of the adductor magnus opening could not be obtained satisfactorily in the majority of cases and this short segment was felt to be a 'blind spot' with the ultrasound technique. The proximal portions of the deep veins of the calf could be seen in only a few cases.

In our study ultrasound demonstrated a 94% sensi- tivity and 100% specificity using contrast venography as the standard. In the one false negative scan an isolated 3cm thrombus was missed in the femoral vein. This was subsequently visualised when the pa- tient was re-scanned in the erect position (Fig. 3). In the series, one case of thrombus confined to the calf

(c)

(d)

Fig. 2 - Femoral vein (arrows) and tributary occluded with thrombus, (a) longitudinal view without compression, (b) longitudi- nal compression view showing little change in calibre, (c) transverse view of the same case showing echogenic thrombus within the femoral vein (arrow) and tributary, (d) longitudinal scan of another case with a long strand of thrombus (thin arrow) within the femoral vein (broad arrows).

REAL-TIME ULTRASOUND DIAGNOSIS OF DEEP VEIN THROMBOSIS 311

(a)

(b)

Fig. 3 - Thrombus (arrow) within the mid-portion of the femoral vein, (a) longitudinal scan, (b) venogram.

veins was detected by venography and in two further cases non-filling of a deep calf vein was considered as possibly representing secondary evidence of thrombus formation. Although we have not found ultrasound useful for detecting isolated calf vein thrombosis, there is evidence that this condition carries a low risk of embolisation (Moser and LeMoine, 1981). Anti- coagulation with its associated morbidity and mortality (Mant et al., 1977) may therefore be inappropriate in this situation.

(a)

(b)

Fig. 4 - (a) Transverse scan of cystic collection in the medial aspect of the popliteal fossa (arrow). Ultrasound showed two other fluid collections in the calf. (b) Arthrogram of the same case showing the popliteal cyst (large arrow) and a second fluid collection posterior to the upper tibia (small arrows). The third fluid collection did not communicate with the joint.

In the single case of an inconclusive ultrasound scan the venogram demonstrated an extremely thin calibre femoral vein with multiple collaterals around the ilio-femoral segment. No thrombus was demonstrated. The appearances were identical to a venogram per- formed 4 years previously and were felt to represent the sequelae of deep vein thrombosis. Although the ultrasound examination demonstrated no thrombus,

312 CLINICAL RADIOLOGY

canal region is not optimal in many cases. Our study shows, however, that real-time ultrasound is capable of detecting deep vein thrombosis in a significant number of cases. It has been suggested that veno- graphy should be performed in all cases of suspected lower limb deep vein thrombosis (Ramsay, 1983). We believe that real-time ultrasound should be considered as an alternative first line investigation.

Fig. 5 - Longitudinal scan showing an aneurysm (open arrow) arising from the distal third of the posterior tibial artery (closed arrows).

owing to the thin calibre and multiplicity of collateral vessels, the examination was reported as inconclusive.

The signs and symptoms associated with deep vein thrombosis may be caused by other pathology for which ultrasound may be diagnostic. In three cases cystic collections in the medial aspect of the popliteal fossa were identified consistent with Baker's cysts (Lukes et al., 1980; Hermann et al., 1981) (Fig. 4). Associated findings in two of these patients also indicated dissection through tissue planes and rupture. Pelvic and inguinal lymphadenopathy was identified in two cases. Percutaneous biopsy subsequently revealed these to be infiltrated with lymphoma and metastatic teratoma. Two cases of haematoma were identified, one with a posterior tibial artery aneurysm (Fig. 5). Both patients had a history of trauma.

Although venography is considered to be the di- agnostic 'gold standard' it also has limitations. If visualisation is inadequate, incorrect therapeutic deci- sions may be made exposing the patient to the side effects of therapy or risks of pulmonary embolism. In our study, venography was impossible in three cases due to failure to cannulate a foot vein. Ultrasound demonstrated patent veins in two of these cases and the presence of thrombus in the third case. Other disadvantages of venography include its inconvenience and possible pain for the patient, the risk of a reaction to contrast media, the development of thrombosis following instillation of contrast media, and the ex- pense of the procedure, particularly if non-ionic con- trast media are used.

Real-time ultrasound, unlike other non-invasive tests, is similar to venography in that it provides direct visual evidence of thrombosis. It is rapid to perform, the average scan time being 5 min, although in cases where thrombus is present the diagnosis may be evident in a matter of seconds. Ultrasound is non- invasive and may detect pathology other than deep vein thrombosis to explain the patient's signs and symptoms. In particular it may obviate the need for combined phlebography and arthrography in the swol- len limb of an arthritic patient, in whom a Baker's cyst and deep vein thrombosis may co-exist (Belch et al., 1981). Ultrasound has the disadvantage of not detect- ing adequately isolated calf vein thrombosis, which may however not be of clinical importance. Visualisa- tion of a short segment of femoral vein in the adductor

Acknowledgement. We are grateful to Drs H. Innes, R. T. MacKay and F. Williams who performed the venograms, and to Mrs P. McQueen for typing the manuscript.

REFERENCES

Albrechtsson, U & Olsson, CG (1976). Thrombotic side effects of lower limb phlebography. Lancet i, 723-724.

Albrechtsson, U & Olsson, CG (1979) Thrombosis following phlebography with ionic and non-ionic contrast media. ACTA Radiologica: Diagnosis (Stockholm), 20 (1), 46-52.

Belch, JJF, McMillan, NC, Fogelman, I, Capell, H & Forbes, CD (1981). Combined phlebography and arthography in patients with painful swollen calf. British Medical Journal, 282, 949.

Browse, N (1978). Diagnosis of deep vein thrombosis. British Medical Bulletin, 34, 163-167.

Effeney, D J, Friedman, MB & Gooding, GAW (1984). Ilio-femoral venous thrombosis: Real-time ultrasound diagnosis: normal criteria and clinical application. Radiology, 150 (3), 787-792.

Flanigan, DP, Goodreau, J J, Burnham, S J, Bergan, JJ & Yao, JS (1978). Vascular-laboratory diagnosis of clinically suspected acute deep vein thrombosis. Lancet, ii, 331-334.

Hermann, G. Yeh, HC, Lehr-Janus, C, Berson, BL (1981). Diagnosis of popliteal cysts: Double contrast arthrography and sonography. American Journal of Radiology, 137, 369-372.

Hull, R, Hirsh, J, Sackett, DL & Stoddart, G (1981). Cost effectiveness of clinical diagnosis, venography and non-invasive testing in patients with symptomatic deep vein thrombosis. New England Journal of Medicine, 304, 1561-1567.

Kristiansen, P, Bergentz, SE, Berggvist, D & Nylander, G (1981). Thrombosis after elective phlebography as demonstrated with the 125-I-fibrinogen test. ACTA Radiologica: Diagnosis (Stock- holm), 22, 577-580.

Lukes, PJ, Herberts, P & Zachrisson, BE (1980). Ultrasound in the diagnosis of popliteal cysts. ACTA Radiologica: Diagnosis (Stockholm), 21 (5), 663-665.

Mant, MJ, O'Brien, BD, Thong, KL, Hammond, GW, Birtwhistle, RV & Grace, MG (1977). Haemorrhagic complications of heparin therapy. Lancet, i, 1133-1135.

Moser, KM, Brach, BB & Dolan, GF (1977). Clinically suspected deep venous thrombosis of the lower extremities: A comparison of venography, impedance plethysmography, and radiolabelled fibrinogen. Journal of the American Medical Association, 237, 2195-2198.

Moser, KM & LeMoine, JR (1981). Is embolic risk conditioned by location of deep venous thrombosis? Annals of Internal Medi- cine, 94, 439-444.

Pochaczevsky, R, Pillari, G Feldman, F (1982). Liquid crystal contact thermography of deep venous thrombosis. American Journal of Radiology, 138 (4), 717-723.

Raghavendra, NB, Rosen, RJ, Lam, S, Riles, T & Horii, SC (1984). Deep venous thrombosis: detection by high resolution real-time ultrasonography. Radiology, 152, 789-793.

Ramsay, LE (1983). Impact of venography on the diagnosis and management of deep vein thrombosis. British Medical Journal, 286, 698--699.

Sandier, DP, Martin, JF, Duncan, JS, Blake, GM, Ward, P, Ramsay, L E e t al. (1984). Diagnosis of deep vein thrombosis: Comparison of clinical evaluation, ultrasound, plethysmography, and venoscan with X Ray venogram Lancet, ii, 716-719.

Schroeder, PJ, Dunn, E (1982). Mechanical plethysmography and Doppler ultrasound diagnosis of deep-venous thrombosis. Archives of Surgery, 117, 300-303.

Simpson, FG, Robinson, PJ, Bark, M & Losowsky, MS (1980).

REAL-TIME ULTRASOUND DIAGNOSIS OF DEEP VEIN THROMBOSIS 313

Prospective study of thrombophlebitis and 'Pseudothrombophle- bitis'. Lancet, i, 331-333.

Sullivan, ED, Peter, DJ & Cranley, JJ (1984). Real-time B-mode venous ultrasound. Journal of Vascular Surgery, i, 465-471.

Thomas, M. Lea (1972). Phlebography. Archives of Surgery, 104, 145-151.

Thomas, M. Lea, Keeling, FP, Piaggio, RB & Treweeke, PS (1984). Contrast agent induced thrombophlebitis following leg phle- bography: Iopamidol versus meglumine iothalamate. British Journal of Radiology, 57, 205-207.

Williams, WJ (1973). Venography (Editorial). Circulation, 47, 220-221.