Thrombosis Research 135 (2015) 673678

Contents lists available at ScienceDirect

Thrombosis Research
journal homepage: www.elsevier.com/locate/thromres

Regular Article

D-dimer for the diagnosis of upper extremity deep and supercial

venous thrombosis
Michelangelo Sartori , Ludovica Migliaccio, Elisabetta Favaretto, Michela Cini,
Cristina Legnani, Gualtiero Palareti, Benilde Cosmi
Department of Angiology and Blood Coagulation, S. Orsola-Malpighi University Hospital, 40138 Bologna, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Background: D-dimer role is well established in the diagnostic work-up for lower limb deep vein thrombosis
Received 10 November 2014 (DVT), however it has not been formally tested for clinically suspected upper extremity DVT and/or supercial
Received in revised form 30 December 2014 vein thrombosis (SVT).
Accepted 7 February 2015 Aim: To ascertain D-dimer diagnostic accuracy for upper extremity DVT and/or SVT.
Available online 14 February 2015
Study design: We performed a single centre management study in outpatients referred by emergency or primary
care physicians for clinically suspected upper extremity DVT. All patients underwent D-dimer testing (cut-off
Keywords:
Upper extremity deep vein thrombosis
value: 500 ng/mL), and a B-mode and color Doppler ultrasonography examination. In case of either technical
Supercial vein thrombosis problems or anatomical barriers, ultrasonography was repeated after 5-7 days. All patients were followed up
D-dimer for three months for the occurrence of symptomatic DVT and/or SVT and/or pulmonary embolism.
Diagnosis Results: We enrolled 239 patients (F: 63.6%; mean SD age: 58.3 16.8). At the initial diagnostic work-up, DVT
Ultrasound imaging was detected in 24 (10%) patients while SVT in 35 (14.6%) patients. During follow-up, one upper extremity DVT
Ultrasonography was found. D-dimer levels were higher in patients with DVT than in those without. Sensitivity and specicity of
D-dimer for DVT were 92% (95%CI: 73-99%) and 60% (95%CI: 52-67%) respectively, with a negative predictive
value of 98% (95%CI: 93-100%), whereas for SVT they were 77% (95%CI: 59-89%) and 60% (95%CI: 52-67%)
respectively, with a negative predictive value of 93% (95%CI: 86-97%).
Conclusions: D-dimer has a negative predictive value 93% for excluding DVT in symptomatic outpatients and it
can be a useful test in the diagnostic work-up of suspected upper extremity DVT.
2015 Elsevier Ltd. All rights reserved.

Introduction with ultrasonography [5]. Such an algorithm safely excluded upper

Upper extremity deep vein thrombosis is not infrequent and accounts
for approximately 14% of cases of deep vein thrombosis (DVT) [1], which
translates into an age-adjusted incidence of 12-19 per 100,000 population
[1]. The risk of pulmonary embolism (PE) in upper extremity DVT is lower
than in lower extremity DVT [1], however patients with upper extremity
DVT may still develop PE [2]. As a result, a prompt diagnosis is important
to select those patients who deserve anticoagulant treatment. D-dimer is
the most frequently used test, in association with ultrasonography, in the
diagnostic work-up for venous thrombo-embolism [3]. D-dimer testing
has been shown to have a high sensitivity and negative predictive value
for both DVT and PE [4]. Recently, D-dimer has been included in a diag-
nostic algorithm for patients with clinically suspected upper extremity
DVT in the largest management study, the ARMOUR, in this setting [5].
D-dimer was used in combination with a clinical decision score and
Corresponding author at: U.O. di Angiologia e Malattie della Coagulazione, Azienda
Ospedaliera di Bologna, Policlinico Sant'Orsola Malpighi, Pad. 2, Via Albertoni, 15, 40138
Bologna. Tel.: +39 051 6362482; fax: +39 051 6362517.
E-mail address: michelangelo.sartori@aosp.bo.it (M. Sartori).
extremity DVT. It has to be noted that in the ARMOUR study the diagnos-
tic accuracy of D-dimer by itself for either upper extremity DVT or upper
extremity supercial vein thrombosis (SVT) was not formally tested. So
far only a small study evaluated D-dimer diagnostic accuracy for upper
extremity DVT in both in- and outpatients [6].
The purpose of this study was to evaluate the diagnostic accuracy of
D-dimer testing for clinically suspected upper extremity DVT and SVT in
out-patients.
Methods
Study Design
This was a single centre prospective management study evaluating
D-dimer accuracy in clinically suspected DVT and/or SVT of the upper
extremity. The study was performed from Jan 1st 2011 to Jun 30th
2013 at a tertiary care teaching hospital (University Hospital S. Orsola-
Malpighi, Bologna, Italy). Symptomatic outpatients referred by general
practitioners or the emergency department to our outpatient clinic
for suspected acute DVT and/or SVT of the upper extremities were

http://dx.doi.org/10.1016/j.thromres.2015.02.007
0049-3848/ 2015 Elsevier Ltd. All rights reserved.
674 M. Sartori et al. / Thrombosis Research 135 (2015) 673678
potentially eligible for the study. Patients were excluded if younger
than 18 years, receiving anti-thrombotic treatment such as low-
molecular -weight heparin or fondaparinux for more than 24 h, vitamin
K antagonists, pregnant or in puerperium, with symptoms attributable to
PE, with life expectancy of less than 3 months. It was not possible to enroll
all consecutive eligible patients but the rst eligible patient of the day was
included, to minimize selection bias. The enrolment was performed dur-
ing business days. At enrolment, each patient underwent: a) medical his-
tory and physical examination; b) D-dimer test; c) ultrasonography.
Patients with clinical signs and symptoms of suspected superior vena
cava thrombosis, (i.e. dyspnea, swelling of face, neck and both arms,
head fullness, cough, chest pain, dysphagia, orthopnea, pleural effusion)
were excluded from the study.
The following risk factors for DVT were assessed using a question-
naire: a) active cancer treatment, ongoing or within previous 6 months
or palliative; b) the presence of a central vein catheter; c) recent insertion
of pacemaker (within last 4 weeks); d) history of vein thrombosis or
previous PE; e) oestrogen-containing therapy; f) a recent peripheral
vein infusion (within one week).
Study outcome was the cumulative 3-month incidence of objectively
conrmed symptomatic upper extremity DVT and PE. Patients with
upper extremity DVT or upper extremity SVT were treated with antico-
agulants, while all other patients were followed up for 3 months to
document the occurrence of symptomatic upper extremity DVT or PE.
In case of worsening symptoms and/or suspected venous thrombo-
embolism, patients were invited to return to our outpatient service:
patients underwent ultrasonography or tomography venography (angi-
ography) in case of suspected upper extremity DVT and helical comput-
ed tomography in case of suspected PE. After three months, all patients
received a phone call or a visit at our outpatient clinic. All outcome events
were adjudicated by one investigator (G.P.) not involved in patient enrol-
ment or follow-up. The study was approved by the local Ethics Commit-
tee. Written informed consent was obtained from all patients.
D-dimer
Blood samples for D-dimer testing were obtained before ultrasonog-
raphy investigation. Blood was drawn by clean venipuncture from an
antecubital vein with a 19-gauge buttery needle and collected into
4 ml plastic tubes containing 0.4 ml 0.106 M trisodium citrate. Whole
blood was centrifuged at 2000 g for 20 min at 20 C. Technicians
performing D-dimer testing were unaware of patient symptoms. The
STA Liatest D-Di (Diagnostica Stago, Asnires, France) was employed
which is an automated and rapid microlatex D-dimer assay. The STA
Liatest D-dimer was performed on the STA Compact coagulation
analyzer as previously described [7]. The results were expressed in
ng/mL (as brinogen equivalent units). As previously described, the cut-
off value for DVT exclusion was 500 ng/mL [7,8].
Ultrasonography Investigation
Patients underwent a comprehensive real-time B-mode and colour
Doppler compression ultrasonography examination of the symptomatic
upper extremity. Ultrasonography investigation was carried out with an
EnVisor C HD instrument (Philips Medical System S.p.A, Monza, Italy),
with a high-resolution broadband width 5-10 MHz linear array trans-
ducer by senior staff vascular medicine physicians who had adequate
experience with color Doppler ultrasonography. The following protocol
was used according to the recommendation of Chin et al.[9]: rstly the
radial, ulnar and brachial veins were scanned in the transverse plane
over their entire length. The diagnosis of DVT was conrmed if there
was lack of compression of the vein, combined with visible intraluminal
thrombus material and/or the absence of venous ow with distal com-
pression. Then the following veins were scanned in the transverse
plane and in the longitudinal plane: axillary, subclavian, and internal
jugular veins. The diagnosis of DVT was conrmed in case of visible
intraluminal thrombus material combined with the absence of venous
ow or an abnormal ow pattern (absent ow or absence of phasic
ow pattern indicating outow obstruction) in the segment of the
vein distal to the thrombosis [9]. Upper extremity SVT was diagnosed
in presence of non-compressibility of the cephalic, basilic, median
antebrachial, median antecubital or accessory cephalic veins [5]. Ultra-
sonographies were considered normal if none of these ndings was
present. If technical problems or anatomical barriers hampered appro-
priate visualization of a portion of the deep veins, the ultrasonography
was considered indeterminate and the procedure was repeated by an
other vascular medicine physician after 5-7 days. If the second ultraso-
nography was inconclusive, the patients underwent a computed tomog-
raphy venography (angiography). The vascular medicine physician was
blind to D-dimer results.
Statistical Analysis
Analysis was carried out using the SPSS software package (version
15.0; SPSS Inc. Chicago, Illinois, USA). Relationships between variables
were assessed using Pearson correlation for continuous variables and
2 or Fisher exact test for categorical variables. Analysis of variance
with Bonferronis correction for multiple comparisons was used to com-
pare means among groups for normally distributed variables. Receiver
operating characteristic (ROC) curves were prepared by plotting the
sensitivity versus 1- specicity and the area under the ROC-curves
(AUC) and the 95% condence interval of AUC for the D-dimer test
were calculated. Categorical variables were expressed as frequencies
and percentages with 95% condence intervals. Continuous variables
were expressed as means SD, inter- quartile ranges (IQR) were also
reported. The signicance level was two-sided and set at = 0.05.
Results
As reported in Fig. 1, 242 patients were screened. Two patients did
not meet the inclusion criteria and were excluded from the study. One
patient had a haematoma and was also excluded. Characteristics of
the enrolled patients (n = 239) are summarized in Table 1. The most
frequent symptoms were pain and oedema, while the most frequent
risk factors for thrombosis were a recent peripheral vein infusion (with-
in one week), active cancer, and history of previous venous thrombo-
embolism (Table 1). No patient with implantable cardioverter debril-
lator or cardiac resynchronization device therapy was enrolled.
At enrolment, ultrasonography revealed 22 (9.2%) upper extremity
DVTs and 35 (14.6%) upper extremity SVTs, but it was inconclusive in
14 (5.9%) patients who underwent repeated testing after 5 to 7 days.
Among the latter, ultrasonography showed no thromboses in 12 patients
but revealed upper extremity DVT in 2 patients (16.7%). No upper
extremity SVT was found among the patients undergoing repeated test-
ing. Since all repeated ultrasonographies were conclusive, no patient
underwent computed tomography venography. Upper extremity DVT
prevalence after the complete diagnostic work-up was 10.0% (n = 24).
All the 180 patients without upper extremity DVT and upper extremity
SVT were followed up for three months, and one patient developed
upper extremity DVT (3-month venous thrombo-embolic incidence
rate: 0.56% (CI95%: 0.01-3.09%). Since patients with negative ultrasound
did not undergo angiography, we were not able to establish whether
such upper extremity DVT was a diagnostic failure or an intervening com-
plication. For further analysis, we considered the additional DVT case that
subsequently developed during the follow-up. No patient developed PE.
At the end of follow-up, the cumulative upper extremity DVT prevalence
was 10.5% (n = 25).
Considering only the single test performed on the enrolment day,
ultranosonography could have missed 3 DVTs (failure rate 1.3%, 95%
CI: 0.46-3.61%). The failure rate of the diagnostic work-up, which
included a second examination with ultrasonography after 5 to 7 days,
was 0.42% (95% CI: 0.02-2.30%).
 M. Sartori et al. / Thrombosis Research 135 (2015) 673678 675

Screened = 242

Clinical assesment

Haemathoma (excluded) On anticoagulant therapy

n=1 (excluded) n=2

D-dimer n=239

B-mode and colour Doppler n=239

No DVT inconclusive DVT SVT

n= 168 n= 14 n= 22 n=35

B-mode and colour Doppler

after 5-7 days

No DVT DVT
n= 12 n= 2

3-month follow-up
DVT n=1

Fig. 1. Study ow chart. SVT, upper extremity supercial vein thrombosis; DVT, upper extremity deep venous thrombosis.

As shown in Table 2, patients with upper extremity SVT were youn-

Table 1 ger compared to those with DVT and to those without thrombosis
Characteristics of the study population (n = 239).
(Table 2). Risk factors for venous thrombo-embolism were the presence
Age SD (IQR): 58.3 16.8 years (24.6) of central venous catheter and a recent pace maker insertion, which
Male/female (%): 87/152 (63.6) were more prevalent in patients with upper extremity DVT than in
Active cancer (%): 39 (16.3)
those without thrombosis and in those with SVT (Table 2). Patients
CVC (%): 14 (5.9)
PM (%): 4 (1.7) with upper extremity DVT had oedema of the symptomatic arm more
History of vein thrombosis (%): 25 (10.5) frequently than those without DVT. Patients with upper extremity SVT
Oestrogen-containing therapy (%): 7 (2.9) had redness or rash more frequently than those with DVT and than
Peripheral vein infusion (%) 44 (18.4)
those without thrombosis and had oedema less frequently than those
Symptoms
with DVT and than those without thrombosis (Table 2). A recent periph-
Pain (%): 173 (72.4) eral vein infusion was more prevalent in patients with SVT than in those
Oedema (%): 129 (54.0) with DVT or in those without thrombosis (Table 2).
Redness or rash (%): 60 (25.1) D-dimer plasma levels were higher in patients with upper ex-
CVC central vein catheter; PM, pacemaker; Anticonception or hormone replacement. tremity DVT than in those without thrombosis (2917 4872 vs.
676 M. Sartori et al. / Thrombosis Research 135 (2015) 673678

Table 2
Characteristic of the study population according to the presence of upper extremity deep vein thrombosis and supercial vein thrombosis.

C SVT DVT C vs. SVT C vs. DVT DVT vs. SVT

n = 179 n = 35 n = 25 p-value p-value p-value p-value

Age (years) SD: 58.8 16.9 51.8 14.5 64.0 16.8 0.014 0.069 0.400 0.014
Male: 33.5% 48.6% 42.3% 0.221
D-dimer (ng/mL) SD 677 797 1838 2940 2917 4872 0.001 0.010 0.001 0.140

Venous thromboembolism risk factors (%)

Active cancer: 18.3 14.7 7.7 0.411
CVC: 4.0 2.9 24 0.001 0.808 0.001 0.050
PM: 1.1 0 8.0 0.030 0.889 0.013 0.056
History of vein thrombosis: 10.6 8.6 12.0 0.904
Oestrogen-containing therapy: 2.5 3.0 7.7 0.348
Peripheral vein infusion 12.1 61.8 7.7 0.0001 0.001 0.604 0.001

Symptoms (%)
Pain: 72.5 93.8 79.2 0.033 0.026 0.374 0.179
Oedema: 58.2 31.3 84.0 0.001 0.17 0.025 0.001
Redness or rash: 18.8 62.5 32.0 0.001 0.001 0.187 0.007

C, patients without thrombosis; SVT, patients with upper extremity supercial vein thrombosis; DVT, patients with upper extremity deep vein thrombosis; CVC central vein catheter; PM,
pacemaker. Statistical analysis was performed by analysis of variance with Bonferronis correction for multiple comparisons for age and D-dimer, and chi-square test for the other variables.
p-value was calculated with multiple logistic regression analysis with correction for age.

677 797 ng/mL, p = 0.001) and they were higher in patients with
upper extremity SVT than in those without thrombosis (1838 2940
vs. 677 797 ng/mL, p = 0.01). Assuming a 500 ng/mL cut-off value
as suggested by the studies on lower limb DVT [4,7], D-dimer was
negative in 107 (59.8%) patients without upper extremity thrombosis,
in 8 (22.9%) patients with upper extremity SVT, and in two with
upper extremity DVT (8.0%). Among these two patients with D-dimer
below the cut-off and with DVT, one was positive at repeated ultraso-
nography during initial diagnostic work-up and the other was the
patient that developed upper extremity DVT during follow-up. Sensitiv-
ity and specicity of D-dimer for upper extremity SVT were 77% (95%CI:
59-89%) and 60% (95%CI: 52-67%) respectively, with a negative predic-
tive value of 93% (95%CI: 86-97%) and a positive predictive value of 27%
(95%CI: 19-37%). Sensitivity and specicity of D-dimer for upper
extremity DVT were 92% (95%CI: 73-99%) and 60% (95%CI: 52-67%)
respectively, with a negative predictive value of 98% (95%CI: 93-100%)
and a positive predictive value of 25% (95%CI: 17-35%). The receiver
operating characteristic (ROC) curve for D-dimer test for upper extrem-
ity DVT is shown in Fig. 2 with an AUC of 0.81 (95%CI: 0.74 -0.89,
p b 0.001).
value of 100% (95%CI: 47100%) in a sample of 52 consecutive patients.
However, the authors concluded that D-dimer was not useful since only
10% of their patient population had normal D-dimer. It has to be noted
that the study of Marminod et al. included 23 patients with cancer and
18 patients with central venous catheters in a total sample of 52 patients,
most of them being inpatients [6]. According to such preliminary report,
Grant et al. [14] and Bernardi et al. [15] did not recommend D-dimer test-
ing in their reviews about upper extremity DVT, because most patients
with suspected upper extremity DVT may have increased D-dimer levels
owing to comorbidities, recent procedures, or indwelling central venous
catheters. We demonstrated that D-dimer is negative in almost half of
the outpatients with suspected upper extremity DVT and has a high

Discussion

In the present study we report D-dimer sensitivity and specicity for

the diagnosis of DVT and of SVT of the upper extremity. The negative
predictive value of D-Dimer was 98% (95%CI: 93100%) for upper
extremity DVT and 93% (95%CI: 86-97%) for upper extremity SVT. More-
over in a diagnostic work-up that included repeated ultrasonography,
only 1 out of 180 patients with a normal ultrasound at presentation
developed DVT during the 3 month follow-up. The role of D-dimer
assays in the diagnosis of patients with suspected deep vein thrombosis
of the legs has been extensively studied. D-dimer assay has proven to be
a highly sensitive but non-specic test for the presence of venous
thrombo-embolism [4] and to have a high negative predictive value for
DVT in different tested patient populations [1013]. A meta-analysis
showed that all D-dimer assays had high sensitivity for proximal DVT of
the lower extremities [4]. Recently we showed that D-dimer had sensitiv-
ity (84%) and negative predictive value (N93%) similar to that found in the
aforementioned meta-analysis also for isolated calf vein thrombosis [7]. In
contrast to lower limb DVT, only one small study investigated the diag- Fig. 2. Receiver Operating Characteristic (ROC) curve analysis of accuracies of D-dimer test for
nostic accuracy of D-dimer in upper extremity DVT: Marminod et al. [6] the presence of upper extremity deep venous thrombosis. * Indicates cut-off value of
reported a sensitivity of 100% (95%CI: 78100%) and a negative predictive 500 ng/mL.
 M. Sartori et al. / Thrombosis Research 135 (2015) 673678
negative predictive value for DVT. Thus, D-dimer can be a useful test in the
diagnostic work-up for suspected upper extremity DVT. In fact, D-dimer
has been included in a new diagnostic algorithm for patients with clinical-
ly suspected upper extremity DVT in the ARMOUR study [5]. In patients
with low probability of upper extremity DVT, such algorithm was able
to safely exclude DVT without the use of the ultrasonography. It has to
be noted that the study of Kleinjan et al. did not formally test D-dimer di-
agnostic accuracy by itself [5]. Moreover, Kleinjan et al. considered both
SVT and DVT as upper extremity DVT in their diagnostic algorithm [5].
In our opinion, pooling together SVT and DVT raises some concerns. SVT
diagnosis is based on the clinical signs and symptoms that include all
components of inammation (pain, reddening of the skin and swelling
of the surrounding tissue, leading to a warm, tender and swollen area or
cord along a supercial vein) [16] and these are not included in the
Constans Clinical Decision Score [17] used in the aforementioned algo-
rithm [5]. In the present study, inammation signs and symptoms were
peculiar for SVT, whereas oedema for DVT. Moreover, pooling together
SVT and DVT implies that D-dimer has a similar diagnostic accuracy for
SVT and DVT. Our study demonstrated that a negative D-dimer test is
less reliable in patients with suspected upper extremity SVT than in pa-
tients with suspected DVT, being negative in more than 20% of patients
with SVT. Recently, we found that D-dimer plasma levels correlated
with the thrombotic burden in patients with lower limb DVT [18] and a
normal D-dimer in upper extremity SVT can be ascribed to the low
thrombotic burden associated with SVT. Our series support the use of ul-
trasound examination in case of clinically suspected SVT for a correct di-
agnosis also in patients with normal D-dimer. In accordance with the
study of Kleinjan et al. [5], we further support the value of D-dimer test
in the diagnostic algorithms for clinically suspected upper extremity
DVT, for its high negative predictive value (98%).
For upper extremity thrombosis diagnosis, we used ultrasonogra-
phy, which is not the gold standard for thrombosis of the upper limb.
However, contrast venography, the gold standard for upper extremity
DVT, is an invasive test requiring ionizing radiations and with several
potential adverse events such as contrast allergies and renal dysfunc-
tion. The non-feasibility of venography in a large number of patients
in our setting compelled us to use the duplex ultrasonography. Com-
pression and Doppler ultrasonography, used alone or in combination,
are the most frequently used objective tests for upper extremity DVT
diagnosis [19]. A meta-analysis estimated that the sensitivity of duplex
ultrasonography for upper extremity DVT ranged from 56% to 100%,
and the specicity ranged from 94% to 100% [20], but no prospective
management study withholding anticoagulant therapy only on the
basis of normal ultrasonography was included. To overcome all the
aforementioned aws of studies on ultrasonography, we performed a
prospective study with a 3 month follow-up. We found that a single
ultrasonography was inconclusive in about 6% of the patients and the
prevalence of upper extremity DVT at the second examination was
not low (16.7%), suggesting that a single examination may not be
sufcient to exclude upper extremity DVT. During follow-up, one
patient developed upper extremity DVT suggesting a 0.4% failure rate
of our diagnostic work-up with repeated ultrasonography. Such a gure
is not different from the 3-month venous thrombo-embolic incidence in
management studies for lower limb DVT [21].
In our series DVT prevalence was approximately 10%, and it was
lower than that observed in the recent ARMOUR study [5]. It has to
be noted that Kleinjan et al. [5] enrolled both in- and outpatients
with a prevalence of 35% central venous catheters, whereas we
enrolled only outpatients with a much lower prevalence of central
venous catheters or pacemakers. Our data shows that the prevalence
of upper extremity DVT is increased in patients with central vein
catheters when compared to patients without. As a result, the
lower DVT prevalence in comparison with the aforementioned
study is not unexpected. SVT prevalence in our series was 14.6%
and it also reects a higher number of subjects with peripheral
vein infusion thrombophlebitis.
677
Some limitations of the present study should be acknowledged. The
number of patients with DVT was rather limited and we did not track
any previous hospitalization (for medical or surgical purposes) within
90 days, that is considered an important risk factor for upper extremity
DVT. The patients investigated were not all consecutive patients refer-
ring to our vascular outpatient service and a selection bias cannot be
excluded. However, with the aim of reducing this bias, most of the
patients included were eligible subjects who presented rst in the
morning. Moreover, the diagnostic accuracy of ultrasonography for
suspected upper extremity DVT is still uncertain. It has been evaluated
in few studies, often with major methodological limitations [20]. We
tried to develop a clear protocol for B-mode and colour Doppler ultraso-
nography examination of the arm. Compression ultrasonography
requires direct manual compression, which cannot be applied to the
medial segment of the subclavian vein underlying the bony clavicle.
Thus a B-mode and colour-ow Doppler ultrasonography was per-
formed for axillary, subclavian, and internal jugular veins. The B-mode
imaging with the addition of colour allowed the visualization of such
venous vessels. Moreover the alterations of the normal biphasic pattern
on pulsed-wave Doppler ow analysis also suggested DVT presence
[22]. The systematic review by di Nisio et al. [20] evaluated nine studies
using ultrasonography to diagnose upper extremity DVT and the
estimated sensitivity and specicity were 91% (95%CI: 8597%) and
93% (95%CI: 80100%) for Doppler ultrasonography with compression.
Since the accuracy of ultrasonography was not high enough for use as
a gold standard, we designed a prospective study and also evaluated
thrombo-embolic event rate at three months. It has to be noted that
so far, no management study had been performed on the safety of with-
holding anticoagulation in suspected DVT only relying on negative
ultrasonography. The use of Doppler ultrasonography imaging may
represent a method of detection of the stenosis and vascular compres-
sion associated with thoracic outlet syndrome that is associated with
subclavian-axillary vein thrombosis [23]. For such diagnosis Doppler
ultrasonography is associated with both high-sensitivity and high-
specicity particularly when combined with the Adson's test [24]. How-
ever, in our vascular emergency room, we do not routinely perform any
search for the thoracic outlet syndrome. Finally we did not use a strict
clinical strategy (using a predictive score such as the Costans score),
so a single D-dimer test having a sensitivity of 92% may be fallaciously
reassuring.
In conclusion, our results indicate that the negative predictive value
of a normal D-Dimer was N93% for upper extremity DVT, and D-dimer
seems a useful test in the diagnostic work-up for suspected upper
extremity DVT.
Funding
None.
Disclosure of Conict of Interests
The authors state that they have no conict of interest.
References
[1] Spencer FA, Emery C, Lessard D, Goldberg RJ, Worcester Venous Thromboembolism
Study. Upper extremity deep vein thrombosis: a community-based perspective. Am
J Med 2007;120:67884.
[2] Munoz FJ, Mismetti P, Poggio R, Valle R, Barrn M, Guil M, et al. Clinical outcome
of patients with upper-extremity deep vein thrombosis: results from the RIETE
Registry. Chest 2008;133:1438.
[3] Prisco D, Grifoni E. The role of D-dimer testing in patients with suspected venous
thromboembolism. Semin Thromb Hemost 2009;35:509.
[4] Goodacre S, Sampson FC, Sutton AJ, Mason S, Morris F. Variation in the diagnostic
performance of D-dimer for suspected deep vein thrombosis. QJM 2005;98:51327.
[5] Kleinjan A, Di Nisio M, Beyer-Westendorf J, Camporese G, Cosmi B, Ghirarduzzi A,
et al. Safety and feasibility of a diagnostic algorithm combining clinical probability,
d-dimer testing, and ultrasonography for suspected upper extremity deep venous
thrombosis: a prospective management study. Ann Intern Med 2014;160:4517.
678 M. Sartori et al. / Thrombosis Research 135 (2015) 673678
[6] Merminod T, Pellicciotta S, Bounameaux H. Limited usefulness of D-dimer in
suspected deep vein thrombosis of the upper extremities. Blood Coagul Fibrinolysis
2006;17:2257.
[7] Sartori M, Cosmi B, Legnani C, Favaretto E, Valdr L, Guazzaloca G, et al. The Wells
rule and D-dimer for the diagnosis of isolated distal deep vein thrombosis. J Thromb
Haemost 2012;10:22649.
[8] Sidelmann JJ, Gram J, Larsen A, Overgaard K, Jespersen J. Analytical and clinical
validation of a new point-of-care testing system for determination of D-dimer in
human blood. Thromb Res 2010;126:52430.
[9] Chin EE, Zimmerman PT, Grant EG. Sonographic evaluation of upper extremity deep
venous thrombosis. J Ultrasound Med 2005;24:82938.
[10] Scarano L, Bernardi E, Prandoni P, Sardella C, Rossi L, Carraro P, et al. Accuracy of two
newly described D-dimer tests in patients with suspected deep venous thrombosis.
Thromb Res 1997;86:939.
[11] Leroyer C, Bressollette L, Oger E, Le Ber C, Nonent M, Boulch MT, et al. Diagnostic
value of plasma D-dimer measurement, using ELISA test, in front of a suspected
deep venous thrombosis. Eur J Intern Med 1996;7:99103.
[12] Wells PS, Brill-Edwards P, Stevens P, Panju A, Patel A, Douketis J, et al. A novel and
rapid whole-blood assay for D-dimer in patients with clinically suspected deep
vein thrombosis. Circulation 1995;91:21847.
[13] Lindahl TL, Lundahl TH, Ranby M, Fransson SG. Clinical evaluation of a diagnostic
strategy for deep venous thrombosis with exclusion by low plasma levels of brin
degradation product D-dimer. Scand J Clin Lab Invest 1998;58:30716.
[14] Grant JD, Stevens SM, Woller SC, Lee EW, Kee ST, Liu DM, et al. Diagnosis and manage-
ment of upper extremity deep-vein thrombosis in adults. Thromb Haemost 2012;108:
1097108.
[15] Bernardi E, Pesavento R, Prandoni P. Upper extremity deep venous thrombosis.
Semin Thromb Hemost 2006;32:72936.
[16] Bauersachs RM. Diagnosis and treatment of supercial vein thrombosis.
Hamostaseologie 2013;33:23240.
[17] Constans J, Salmi L-R, Sevestre-Pietri M-A, Perusat S, Nguon M, Degeilh M, et al. A
clinical prediction score for upper extremity deep venous thrombosis. Thromb
Haemost 2008;99:2027.
[18] Sartori M, Favaretto E, Cini M, Legnani C, Palareti G, Cosmi B. D-dimer, FVIII and
thrombotic burden in the acute phase of deep vein thrombosis in relation to the
risk of post-thrombotic syndrome. Thromb Res 2014;134:3205.
[19] Mustafa BO, Rathbun SW, Whitsett TL, Raskob GE. Sensitivity and specicity of ultra-
sonography in the diagnosis of upper extremity deep vein thrombosis a systematic
review. Arch Intern Med 2002;162:4014.
[20] Di Nisio M, van Sluis GL, Bossuyt PMM, Buller HR, Porreca E, Rutjes AWS. Accuracy of
diagnostic tests for clinically suspected upper extremity deep vein thrombosis: a
systematic review. J Thromb Haemost 2010;8:68492.
[21] Bernardi E, Camporese G, Bller HR, Siragusa S, Imberti D, Berchio A, et al. Serial
2-point ultrasonography plus D-dimer vs whole-leg color-coded Doppler ultraso-
nography for diagnosing suspected symptomatic deep vein thrombosis: a random-
ized controlled trial. JAMA 2008;300:16539.
[22] Patel MC, Berman LH, Moss HA, McPherson SJ. Subclavian and internal jugular veins
at Doppler US: abnormal cardiac pulsatility and respiratory phasicity as a predictor
of complete central occlusion. Radiology 1999;211:57983.
[23] Nichols AW. Diagnosis and management of thoracic outlet syndrome. Curr Sports
Med Rep 2009;8:2409.
[24] Lee AD, Agarwal S, Sadhu D. Doppler Adson's test: predictor of outcome of surgery in
non-specic thoracic outlet syndrome. World J Surg 2006;30:2912.