**6. Diagnostic procedures**

As in thrombosis of the lower extremities, there are no reliable clinical symptoms to diagnose deep arm vein thrombosis. Therefore, diagnostic test are necessary to diagnose or rule out upper extremity thrombosis. Contrast venography is the gold standard diagnostic method, allowing an unparalleled overview of all arm veins with high resolution (Fig 2). However, venography is invasive, inconvenient for the patient and suffers from moderate inter-observer agreement rates (between 71 and 83%) in upper extremity deep vein thrombosis (Baarslag et al*.*, 2003). Furthermore, possible complications as contrast agent mediated kidney damage, allergic reactions and even venography-induced thrombosis can occur (Bernardi et al*.*, 2006).

Ultrasound has several advantages compared to invasive methods and has become the diagnostic method of choice in lower extremity deep vein thrombosis (Kearon et al*.*, 1998; Goodacre et al*.*, 2006). The main strengths of ultrasound are its non-invasive nature, general availability, as well as the lack of radiation and contrast material. However, ultrasound suffers generally from some disadvantages, including observer variability and usually lack of standardized documentation. In arm vein thrombosis, an additional obstacle is the portion of the subclavian vein behind the clavicle, impeding compression manoeuvres in this clinically important region. In addition, the brachiocephalic veins and the superior vena

Deep Vein Thrombosis of the Arms 135

Fig. 3. Computed tomography scan (A, Department of Radiology, Innsbruck Medical University) and colour duplex ultrasound (B) of the same patient showing a floating

CUS 96% 94% noninvasive operator-

CUS + CD 100% 93% noninvasive operator-

**A B**

**Test Sensitivity Specificity Advantage Disadvantage Reference** Clinical score 78% 64% simplicity performance Constans et al. 2008 D-Dimer 100% 14% sensitivity specficity Merminod et al. 2006

MRI 71% 89% overview cost Baarslag et al. 2004 Venography 100% 100% overview invasive Baarslag et al. 2003

Fig. 4. Indirect sonographic diagnosis of a non-recent thrombosis of the left sublcavian vein with partial recanalisation. Normal venous flow with cardiac and respiratory modulation in

the right subclavian vein (A) compared to linear flow in the left subclavian vein (B).

Table 2. Strengths and weaknesses of diagnostic tests for suspected deep arm vein thrombosis. CUS compression ultrasound. CD colour Doppler ultrasound. MRI magnetic

dependent Di Nisio et al. 2010

dependent Di Nisio et al. 2010

**A B**

thrombus in the left internal jugular vein.

resonance imaging.

cava cannot be examined directly. In a substantial number of cases it will therefore be necessary to rely on indirect signs, as e.g. lack of Doppler signals (Fig. 3B) or characteristic changes in the Doppler flow distal of the occluded segment (Fig. 4). Nevertheless, several studies have shown high sensitivity and specificity of different ultrasound modalities (continuous wave ultrasound, compression ultrasound, colour Doppler ultrasound) in patients with suspected arm vein thrombosis (Prandoni et al*.*, 1997a; B. O. Mustafa et al*.*, 2002; Di Nisio et al*.*, 2010); see Table 2. Despite these impressive numbers, the clinician should bear in mind that all these studies of arm vein thrombosis have been performed in relatively few patients. Therefore, the reported confidence intervals are wide and the safety of withholding therapy in a patient with negative ultrasound has not been proven prospectively in an adequately powered study (B.O. Mustafa et al. 2002). Even with these limitations, ultrasound is a valuable tool for the diagnosis of deep arm vein thrombosis in the hand of an experienced operator, and should be performed in most cases as the first imaging test. For magnetic resonance imaging, a study with 44 patients comparing time of flight and Gadolinium-enhanced imaging reported a moderate sensitivity and specificity (Baarslag et al*.*, 2004). Computed tomography scanning for the diagnosis of deep arm vein thrombosis has only been described in a small case series (Kim et al*.*, 2003), although this modality is often used to detect thrombi in clinical practice (Fig. 3A).

Fig. 2. Venogram of a chronic occlusion of the left subclavian vein with extensive collateralization (Department of Radiology, Innsbruck Medical University).

Since there is no imaging method combining optimal accuracy and minimal burden for the patient, alternative methods have been searched for. An interesting clinical prediction score for arm vein thrombosis, reminiscent of the Wells score for lower extremity thrombosis (Wells et al*.*, 1997), has been published by Constans and coworkers (Constans et al*.*, 2008). This simple score assigns one point each for a central venous catheter or pacemaker lead, localized pain and unilateral edema. One point is subtracted in the case that an alternative diagnosis would seem at least as likely as deep arm vein thrombosis. D-dimer testing has also been evaluated in a cohort of patients with suspected deep arm vein thrombosis (Merminod et al*.*, 2006). Although nearly 100% sensitive, D-dimer suffers from a low specificity.

cava cannot be examined directly. In a substantial number of cases it will therefore be necessary to rely on indirect signs, as e.g. lack of Doppler signals (Fig. 3B) or characteristic changes in the Doppler flow distal of the occluded segment (Fig. 4). Nevertheless, several studies have shown high sensitivity and specificity of different ultrasound modalities (continuous wave ultrasound, compression ultrasound, colour Doppler ultrasound) in patients with suspected arm vein thrombosis (Prandoni et al*.*, 1997a; B. O. Mustafa et al*.*, 2002; Di Nisio et al*.*, 2010); see Table 2. Despite these impressive numbers, the clinician should bear in mind that all these studies of arm vein thrombosis have been performed in relatively few patients. Therefore, the reported confidence intervals are wide and the safety of withholding therapy in a patient with negative ultrasound has not been proven prospectively in an adequately powered study (B.O. Mustafa et al. 2002). Even with these limitations, ultrasound is a valuable tool for the diagnosis of deep arm vein thrombosis in the hand of an experienced operator, and should be performed in most cases as the first imaging test. For magnetic resonance imaging, a study with 44 patients comparing time of flight and Gadolinium-enhanced imaging reported a moderate sensitivity and specificity (Baarslag et al*.*, 2004). Computed tomography scanning for the diagnosis of deep arm vein thrombosis has only been described in a small case series (Kim et al*.*, 2003), although this

modality is often used to detect thrombi in clinical practice (Fig. 3A).

Fig. 2. Venogram of a chronic occlusion of the left subclavian vein with extensive collateralization (Department of Radiology, Innsbruck Medical University).

specificity.

Since there is no imaging method combining optimal accuracy and minimal burden for the patient, alternative methods have been searched for. An interesting clinical prediction score for arm vein thrombosis, reminiscent of the Wells score for lower extremity thrombosis (Wells et al*.*, 1997), has been published by Constans and coworkers (Constans et al*.*, 2008). This simple score assigns one point each for a central venous catheter or pacemaker lead, localized pain and unilateral edema. One point is subtracted in the case that an alternative diagnosis would seem at least as likely as deep arm vein thrombosis. D-dimer testing has also been evaluated in a cohort of patients with suspected deep arm vein thrombosis (Merminod et al*.*, 2006). Although nearly 100% sensitive, D-dimer suffers from a low

Fig. 3. Computed tomography scan (A, Department of Radiology, Innsbruck Medical University) and colour duplex ultrasound (B) of the same patient showing a floating thrombus in the left internal jugular vein.


Table 2. Strengths and weaknesses of diagnostic tests for suspected deep arm vein thrombosis. CUS compression ultrasound. CD colour Doppler ultrasound. MRI magnetic resonance imaging.

Fig. 4. Indirect sonographic diagnosis of a non-recent thrombosis of the left sublcavian vein with partial recanalisation. Normal venous flow with cardiac and respiratory modulation in the right subclavian vein (A) compared to linear flow in the left subclavian vein (B).

Deep Vein Thrombosis of the Arms 137

anticoagulant therapy in patients with arm vein thrombosis. There is no specific recommendation in the ACCP guidelines on the treatment of cancer patients with upper extremity deep vein thrombosis. In cancer patients with deep arm vein thrombosis, the use of low molecular weight heparins instead of vitamin K antagonist as long-term treatment has been suggested in analogy to lower extremity thrombosis, but there are currently no studies supporting this approach (Shivakumar et al*.*, 2009). Although various degrees of post-thrombotic syndrome have to be expected in the long term follow up of about 1 in 4 patients with upper extremity deep vein thrombosis, the ACCP guidelines do not advocate the routine use of elastic bandages or compression sleeves for the arm, unless patients report

A number of studies have described case series of deep arm vein thrombosis treated with a variety of invasive therapeutic options, including catheter-guided thrombolysis, percutaneous angioplasty with or without venous stent insertion, surgical thrombectomy and surgical decompression of costoclavicular narrowing to correct thoracic inlet syndrome, e.g. by first rib resection (Zimmermann et al*.*, 1981; Becker et al*.*, 1983; Machleder, 1993; Urschel & Razzuk, 1998). Some investigators recommend such an invasive approach routinely e.g in patients with effort related thrombosis (Paget von Schroetter´s syndrome) (Kommareddy et al*.*, 2002). Here, the ACCP guidelines clearly do not recommend invasive procedures routinely, but only in selected patients and in specially equipped centers. It remains to be determined in adequately designed, randomized clinical trials whether these invasive procedures, which carry a substantial risk of major bleeding and other serious complications, provide a benefit compared to standard anticoagulation with optimal

**Subjective symptoms Objective Signs Subjective symptoms Objective Signs**  Heaviness Edema Heaviness Edema

Physical limitation Venous ectasia Cramps Dependent cyanosis Paraesthesia Redness Paraesthesia Redness

Another point of debate is the question whether central venous catheters should be removed when a diagnosis of deep vein thrombosis has been confirmed in the respective vessel. Most experts opt against catheter removal, if the catheter is still needed and still functional. In a cohort study of 74 cancer patients with acute upper extremity thrombosis, the catheters were not removed and patients were treated for 3 months with standard anticoagulation without recurrent episodes of venous thromboembolism (Kovacs et al., 2007). If the catheter is removed, the ACCP guidelines recommend not to shorten the anticoagulation period below

Pain Skin induration Pain Prominent veins on arm

Prominent veins over shoulder or anterior chest wall

Tenderness

severe symptoms like persistent edema and pain.

mechanical compression using elastic bandages.

3 months (Kearon et al*.*, 2008).

Pruritus Discoloration Pruritus

Pain during compression

Prandoni et al. 2004, Vik et al. 2009 Kahn et al. 2005 Table 4. Two suggested modifications of the Villalta scale for the assessment of postthrombotic syndrome in deep arm vein thrombosis. Each sign or symptom is graded as 0 (absent), 1 (mild), 2 (moderate) or 3 (severe). A score of 5 or higher is classified as postthrombotic syndrome and score of 15 or higher as severe post-thrombotic syndrome.
