Variations in Lower Limb

Radiology

Daniel J. Quinlan, MBBS

Raza Alikhan, BSc, MBBS
Philip Gishen, MB, FRCR Venous Anatomy: Implications
Paul S. Sidhu, MB, FRCR
for US Diagnosis of Deep Vein
Index terms:
Veins, anatomy, 93.92
Thrombosis1
Veins, thrombosis, 93.751
Venography, 93.124
PURPOSE: To retrospectively review bilateral venograms free of thrombus to eval-
Published online before print
10.1148/radiol.2282020411 uate the frequency and types of variations seen in venous anatomy.
Radiology 2003; 228:443– 448
MATERIALS AND METHODS: A retrospective review of 404 bilateral (808 limbs)
Abbreviations: lower limb venograms obtained from medical patients participating in a thrombo-
DVT ⫽ deep vein thrombosis prophylaxis clinical trial and found to be free of thrombus was performed.
SFV ⫽ superficial femoral vein Venograms were evaluated according to predetermined criteria for the presence of
duplication of vessels and inter- and intraindividual variations in venous anatomy.
1
From the Department of Radiology, Variations were assessed with analysis of variance and ␹2 tests.
King’s College Hospital, Denmark Hill,
London SE5 9RS, England (D.J.Q., RESULTS: Two vessels were seen in the popliteal fossa on 337 (42%) of 808
P.G., P.S.S.); and Academic Depart- venograms, and 41 (5%) were true duplicated popliteal veins. There were 253
ment of Surgery, Guy’s King’s and St
Thomas’ School of Medicine, London,
(31%) duplicated superficial femoral veins (SFVs), with 12 (1.5%) being complex
England (R.A.) From the 2000 RSNA duplicated systems. Of 265 duplicated SFVs, 138 (52%) began in the midthigh
scientific assembly. Received April region and 80 (30%), in the adductor canal region. The duplicated vessel was
5, 2002; revision requested June 12; medial to the main SFV in 122 (46%), lateral in 131 (49%), and both (ie, triplica-
revision received November 10;
accepted December 10. Address tions) in 12 (4.5%). The length of the duplicated SFV ranged from 1 to 35 cm; 6 –15
correspondence to D.J.Q. (e-mail: cm was the most common length in 162 (62%) SFVs. There was no significant
dan.quinlan@consultoberon.com). association between the incidence of anatomic variations and age or sex (P ⬎ .1).
The presence of multiple vessels in one leg was strongly correlated with the prob-
ability of occurrence in the other leg (P ⬍ .001).
CONCLUSION: Variations in lower limb venous anatomy are common and have
important implications for the US diagnosis of deep vein thrombosis.
© RSNA, 2003

Deep vein thrombosis (DVT) is a common medical condition with a wide range of
manifestations that range from an asymptomatic state to a classic symptomatic DVT, with
important sequelae of pulmonary embolism, chronic venous insufficiency, and postphle-
bitic syndrome. Clinical examination is insensitive, and objective tests are required for the
diagnosis (1). Among the diagnostic tools available for patients with symptomatic DVT,
duplex ultrasonography (US) has become the imaging modality of choice (2), increasingly
Author contributions: replacing venography because of its simplicity and high sensitivity and specificity, espe-
Guarantor of integrity of entire study, cially in the femoropopliteal region (3). US, however, is largely operator dependent and
P.S.S.; study concepts and design, less accurate than venography for the diagnosis of calf vein DVT (3). It is important that
D.J.Q., R.A., P.S.S.; literature research, those who perform this procedure and interpret the results have a sound knowledge of the
D.J.Q., R.A., P.S.S.; clinical studies, all
authors; data acquisition, D.J.Q., R.A.,
normal lower limb venous anatomy and the variations.
P.S.S.; data analysis/interpretation, Variations in the anatomy of the lower limb venous system have been studied with use
D.J.Q., R.A.; statistical analysis, R.A.; of cadavers, venography, and US, with differing results. In most traditional anatomy
manuscript preparation, definition of textbooks, the venous system of the lower limb is described as consisting of a continuous
intellectual content, and revision/re-
flow of veins without duplication (4). It has, however, been demonstrated that only one in
view, D.J.Q., R.A., P.S.S.; manuscript
editing and final version approval, all six patients will have this normal venous anatomy (5). The major variation is seen in the
authors. superficial femoral vein (SFV) where, according to previous study findings, between 6%
© RSNA, 2003 and 46% of patients have duplicated or multiple vessels (6 –10). The next most common
variation is with the popliteal vein (4). By having performed a large number of US and

443
 venographic studies in our department, it SFV, common femoral vein, and external Calf Veins
is apparent that many patients have two and common iliac veins were also ac-
All calf (eg, peroneal, anterior and pos-
or more vessels in the popliteal fossa, and quired. All centers used only nonionic
terior tibial veins, gastrocnemius or mus-
a number of these are true duplications iodinated contrast media, with no spe- cular veins) and proximal vessels had to
but many represent a high confluence of cific type stipulated in the protocol ex- be clearly visualized in both limbs for the
Radiology

the tibial veins. cept that the iodine concentration had to venogram to be eligible for inclusion in
With the current reliance on US for the be a minimum of 200 mg/mL and that the study. In particular, we did not eval-
diagnosis of DVT, we believed there was a the injected volume had to be a mini- uate venograms that showed only partial
need for an extensive review of the ve- mum of 60 mL per limb. filling of the calf veins. For each of the
nous anatomy of the lower limb. The Only those bilateral venograms that calf vessels, a record was made of
availability of bilateral lower limb veno-
were evaluable and found to be free of whether they were paired (as is the classic
grams, obtained in a standard manner
thrombus by the clinical trial central ra- appearance), triplicated, or single. The
from asymptomatic medical patients
diology reading committee (consisting of drainage of each calf vessel into other calf
during a recent clinical trial, allowed the
three experienced physicians) were in- or popliteal vessels and the position of
opportunity to assess the true anatomic
cluded and assessed for variations in this confluence in relation to other ves-
variation of lower limb veins (11). There-
anatomy. To be classified as evaluable, sels were recorded. Using this informa-
fore, the purpose of our study was to ret-
the venogram had to demonstrate ade- tion, we classified how the calf vessels
rospectively review bilateral venograms
quate contrast material filling of all the joined to form the trifurcation prior to
that were free of thrombus by evaluating
major veins. Venograms depicting throm- the formation of the popliteal vein. The
the frequency and types of variations
bus were excluded from our study for two popliteal vein is described as a single ves-
seen in venous anatomy.
reasons. First, thrombus may potentially sel formed by the confluence of the an-
cause obstruction, which would result in a terior and posterior tibial veins, often at
MATERIALS AND METHODS the distal border of the popliteus muscle,
failure of contrast material to fill the veins
which become the SFV proximal to the
A retrospective review of bilateral lower and therefore not be visualized. Second,
adductor opening (4). We recorded visu-
limb venograms was performed. The thrombus may cause formation of collat-
alization of the gastrocnemius vein and
venograms were obtained as part of a eral vessels and, as a consequence, poten-
the position of drainage.
multicenter trial of venous thrombopro- tially giving the impression of vessel dupli-
phylaxis that was previously reported as cation. Venography is acceptable for the
Popliteal Vein
the Medical Patients with Enoxaparin, or assessment of the anatomic position of ves-
MEDENOX, trial (11). These bilateral sels as long as adequate filling of the major For the popliteal vein, we recorded the
venograms had been obtained over a 20- vessels is achieved. In view of this, only number of vessels present in the popliteal
month period (December 1996 to July high-quality venograms characterized by fossa by counting the number crossing
1998) for the assessment of thrombus in full and adequate contrast material filling the knee joint space to be able to relate
the venous system of the lower limbs in of all the deep veins were evaluated. this to sonographic visualization during
acutely ill medical patients at day 10 ⫾ 4 a routine US examination. Second, using
following admission to the hospital. the knee joint space as a reference point,
Local ethics committee approval for Venogram Interpretation we recorded the formation of the popli-
the MEDENOX study was obtained in all teal vein (as defined previously) as arising
participating centers (n ⫽ 60), and a de- Two teams of two readers (D.J.Q., P.S.S. either distal to, proximal to, or at the
tailed standard protocol for patient inclu- and R.A., P.G.), each experienced in per- level of the knee joint (Fig 1).
sion was followed. Patients with a history forming and interpreting venographic
of DVT were excluded from the study. All studies, reviewed all of the studies at a SFV
patients provided informed consent to central reading station over a 4-day pe-
participate in the study and to the per- riod. The final interpretation was reached We recorded whether any multiplicity
formance of bilateral venography. Ethics by consensus of each pair of observers, of the SFV had occurred. Care was taken
committee approval or informed consent with a four-reviewer consensus with re- to identify the deep femoral and long
was not required for the retrospective re- spect to any particularly difficult inter- saphenous veins and exclude them from
view of images, as the images were free of the analysis. The vessels were recorded as
pretation. Prior to commencement of the
patient identification. single, double, or complex. Duplicated
study, a defined protocol was established
vessels were assessed for their position,
according to anatomic definitions used
length, and size with respect to the orig-
Venography in previous studies to correctly and con-
inal SFV, which was defined as the vessel
sistently define the vessels. The observers
All venograms, although obtained at that most closely followed the course of
different centers, were obtained in a stan- underwent a period of training to de- the superficial femoral artery. The posi-
dardized manner, as defined by the pro- velop a uniform standard definition of tion of the duplicated SFVs was recorded
tocol, by using the technique described lower limb venous anatomy (8). For de- as being medial or lateral to the true SFV.
by Rabinov and Paulin (12). A distal foot scriptive purposes, proximal implied a The length (in centimeters) of the dupli-
vein was cannulated, and contrast mate- central (cranial or cephalad) location, cated vessel was estimated subjectively
rial was injected to opacify the deep whereas distal implied a peripheral (cau- from the images by calculating the
veins. Images of the calf veins were ob- dal) location. The proximal deep leg length in relation to the estimated femur
tained in three views: two oblique and veins were defined as those central to the length. We divided the length of the SFV
one anteroposterior. Two views of the popliteal vein and the distal veins, as into the following 5–10-cm groups: 1–5
popliteal vein and one view each of the those peripheral to the popliteal vein. cm, 6 –10 cm, 11–20 cm, 21–30 cm, and

444 䡠 Radiology 䡠 August 2003 Quinlan et al

 TABLE 1
Origin of Popliteal Vein and Number
of Vessels in the Popliteal Fossa
Popliteal Vein Right Left Overall*
Radiology

Origin
Knee joint 23 25 48 (5.9)
Proximal 122 109 231 (28.6)
Distal 259 270 529 (65.5)
No. of vessels
Single 222 230 452 (55.9)
Double 172 165 337 (41.7)
Triple 10 9 19 (2.4)
True duplication 21 20 41 (5.1)
* Data in parentheses are percentages.

Figure 1. Illustration demonstrates variations in the formation of

popliteal vein at the knee joint (A), distal to knee joint (B), and into the posterior tibial vein, 32% (243 of
proximal to knee joint (C), as well as true duplication of the popliteal 756) of the time; and into the anterior
vein (D). tibial vein, 8% (63 of 756) of the time.
Gastrocnemius veins were only visual-
ized 64% (432 of 808) of the time, and in
76% (315 of 414) of cases the drainage
31 cm or greater. For a percentage size, an anomalies was tested by using analysis of was above the knee joint. A variable
estimation of the size of the duplicated variance. P values less than .05 were con- number of gastrocnemius veins (one to
vein was made subjectively by dividing sidered to indicate a significant differ- six) were visualized.
the width of the duplicated vein by the ence (or association).
width of the native SFV at the same level. Popliteal Vein
Furthermore, the level at which the du- RESULTS
plication arose, either below the adduc- Data on the popliteal vein are pre-
tor canal, at the adductor canal, or above sented in Table 1. The popliteal vein
We retrospectively reviewed the 404 bi-
the adductor canal, was noted. Record commenced at the knee joint or proxi-
lateral technically accurate venograms
was made of any direct communication mal to it in 279 (35%) and distal to the
(808 limbs) that met our criteria out of a
between SFV and profunda femoris (deep knee joint in 529 (65%) of the 808 limbs.
total of 718 bilateral venograms obtained
femoral) vein through distal anastomo- Within the popliteal fossa, a single vessel
in medical patients participating in the
ses. was identified in 452 (56%) venograms,
MEDENOX clinical trial (11).
and two or more vessels were identified
Statistical Analysis in 356 (44%) of 808 limbs. True embryo-
Demographics
logical duplication of the popliteal vein
The age and sex of the patients were was present in 41 (5%) of 808 limbs (Fig 2).
The mean age of the patients was 73.8
recorded. Variations in the anatomy of
years (age range, 41–93 years). There were
the lower limb veins were assessed cen-
199 (49%) female patients, and there was SFV
trally and recorded in a statistical data-
no statistical difference in the incidence
base (SPSS 9.0 for Windows; SPSS, Chi- Data on the SFV are presented in Ta-
of venous anomalies attributable to age
cago, Ill) and presented in a table form. bles 2 and 3. Multiple SFVs were identi-
or sex (P ⬎ .1). Analysis of variations in
The frequency distribution of multiple fied in 265 (32.5%) of 808 limbs, of
the calf and popliteal and SFVs showed a
vessels was listed according to sex, and which the majority (253 of 265) were
strong correlation between the presence
the mean age with SD was listed accord- duplicated (Fig 3a, 3b). The remaining
of multiple vessels in one leg and the
ing to the number of vessels. To show the 1.5% (12 of 808) systems were more com-
probability of this occurring in the other
correlation between two legs in the pres- plex (eg, triplication or other anomalies)
leg in the same patient (all P ⬍ .001).
ence of multiple vessels, data from two (Fig 3c). A medial duplication was
legs were listed in a paired format. present in 128 (48%) of the 265 dupli-
Calf Veins
The probability of variations in venous cated vessels, with 138 (52%) beginning
anatomy was compared between limbs, The majority of anterior tibial, poste- in the midthigh region (above the adduc-
sex, and age. The ␹2 test was used in the rior tibial, and peroneal veins were tor canal) and another 80 (30%) starting
assessment of the association of sex with paired, with values of 68% (455 of 672), in the adductor canal region. The dupli-
the variation of venous anomalies and 76% (645 of 792), and 76% (689 of 808), cated SFV varied in length between 1 to
the correlation between two legs in the respectively. Single veins were seen in more than 30 cm, with a mean length of
presence of multiple vessels in the same 33% (225 of 672), 17% (132 of 792), and 10 cm ⫾ 6.7 (SD). Of the 265 multiple
patient. When the expected numbers of 6% (50 of 808) of the veins, respectively. SFVs, 122 (46%) were less than a third of
some cells were less than five, the neigh- Three or more peroneal veins were seen the diameter of the main SFV, while 63
bor categories were combined to avoid in 8% (64 of 808) of patients. Drainage of (24%) were half the diameter. In only a
invalid comparison with the ␹2 test. The the peroneal veins into the trifurcation minority (4%) of cases (five of 131) was
association between age and venous occurred 59% (449 of 756) of the time; the duplicated SFV the same size.

Volume 228 䡠 Number 2 Variations in Lower Limb Venous Anatomy 䡠 445

 TABLE 2 TABLE 3
Number of SFVs Position, Length, and Lowest Point
of Duplicated SFV
SFV Right Left Total*
SFV Right Left Overall*
Single 282 261 543 (67.2)
Radiology

Double 118 135 253 (31.3) Position

Complex 4 8 12 (1.5) Medial 57 65 122 (46.1)
Lateral 61 70 131 (49.4)
* Data in parentheses are percentages. Both 4 8 12 (4.5)
Length (cm)
1–5 9 14 23 (9.0)
6–10 55 49 104 (41.0)
11–20 48 27 75 (29.0)
Deep Femoral Vein and Iliac Veins 21–30 8 12 20 (8.0)
⬎30 2 2 4 (1.6)
The deep femoral vein was seen to Lowest point
communicate with the SFV in 31% (252 Below patella 3 3 6 (2.3)
of 808) of cases, with the internal iliac Above patella 18 23 40 (15.5)
Adductor canal 35 45 80 (30.2)
vein being visualized in only 16% (127 of
Above adductor canal 66 72 138 (52.0)
808) of cases.
* Data in parentheses are percentages.
Symmetry between the Limbs
Symmetry between the two sides was
seen 14% (55 of 404) of the time, with nous anomalies (9,10,13) compared with
the main areas of asymmetry being pop- venographic studies (7,8). US will not
liteal veins (32%, 129 of 404) and SFVs produce the same anatomic venous Figure 2. Coronal veno-
(48%, 195 of 404). gram at the level of the
“map” and tends to be a subjective assess- knee demonstrates true
ment that is dependent on operator skill, popliteal vein duplica-
DISCUSSION especially in examinations below the tion that corresponds to
knee (2). For these reasons, a US ana- that in Figure 1, D.
In the present study, we have assessed tomic study of the lower limb veins will
medical patients who were free of DVT in have limitations.
order to assess the true incidence of vari- The majority of pulmonary emboli
ations in venous anatomy among the arise from the proximal deep venous sys- thirds with thrombus in a single SFV,
general population. We have visualized tem of the lower limbs (14,15), and there- which is likely due to the presence of
the whole venous system by using the fore, accurate assessment of this area is of collaterals (7). An additional important
established standard for imaging the particular importance, more so when the point is whether a duplicated SFV is an
lower limb anatomy and have demon- incidence of duplication of the SFV is as independent risk factor for DVT. Further
strated the normality of the distal calf high as 30%, as suggested by findings of studies are required to assess this poten-
venous system and the abnormality of the present study. Our findings are in tial association.
the proximal popliteal and superficial keeping with those of Screaton et al (8), We demonstrated that the majority of
femoral venous system in comparison to who examined 381 venograms and con- duplicated SFVs arise at the adductor ca-
textbook anatomy. cluded that as many as 46% of patients nal or just above it in the midthigh re-
We believe the present study has a have duplicated and/or multiple SFVs. gion, with an equal chance of lying me-
number of strengths compared with This figure was higher than that in previ- dial or lateral to the native vessel. There
those in previous studies (7,8,13). All pa- ous studies, which had shown an inci- was a wide range of variation in the size
tients did not undergo surgery and were dence of 20%–25% (6,9,16). This varia- of the duplicated vein. The adductor ca-
free of thrombus, with visualization of tion may reflect differences in sampling nal is accessible to US over the anterior
the entire deep venous system of both techniques (US or venography) (9), aspect of the thigh, and it should be tech-
lower limbs (8). None of the patients in- smaller sample sizes of earlier studies, nically possible to visualize duplicated
cluded in this study was at high risk of and the fact that many studies reviewed femoral veins when the vessels are as-
developing DVT and, therefore, the sam- patients with a symptomatic DVT (7). sessed in the transverse plane. This will
ple chosen is representative of a normal Duplication of the SFV is a recognized also depend on the transducer footprint
population. Inclusion of high-risk pa- antecedent of missed proximal thrombus length. A linear-array multifrequency
tients in this study could potentially bias at US (7,8,17), with the incidence being probe (15L8w; Acuson Sequoia, Moun-
any results, with a particular anatomic twice as high in patients with a dupli- tain View, Calif) may have a footprint
variation being more prone to produce cated SFV compared with those with a size of up to 5.5 cm, which allows a broad
symptomatic thrombus, which would single SFV (7). The reason for this may view either side of the native SFV. Al-
lead to further investigation. Venogra- result from a decrease in blood flow ve- though in the present study we did not
phy rather than US was used to delineate locity and the subsequent pooling of document the distance of any duplicated
the anatomy, despite its infrequent cur- blood in the duplicated vein, which stim- vein from the native vein, the impression
rent use for initial DVT imaging (2). Al- ulates the formation of thrombus (7,13). was that a large number would be seen in
though US studies have included larger Furthermore, less than half of patients this manner. Size is also an important
numbers of patients, they have consis- with thrombus in a duplicated SFV are factor, with only 4% of duplicated vessels
tently demonstrated lower rates of ve- symptomatic compared with almost two- being the same size as the native vessel;

446 䡠 Radiology 䡠 August 2003 Quinlan et al

 having thrombus extending into the
common femoral or popliteal veins or
both. Thus, assessment of the SFV is of-
ten of great importance in the establish-
ment of more than 20% of all lower limb
Radiology

DVT (7). Therefore, like other authors

(25,27), we cannot support the sugges-
tion that it is safe to exclude SFV in US
assessment of the lower limb.
Our study has several potential limita-
tions. First, by selecting high-quality
venograms with full opacification, we
may have missed some variations that
lend themselves to poor contrast material
filling. This may be of importance in
multiple SFVs, some of which might not
have filled with contrast material. Sec-
ond, in a review of this nature, where
numerous investigators in different
countries have performed examinations,
individual venographic techniques may
have influenced outcomes. Third, our
calculations of SFV length should be re-
garded as estimates, but we attempted to
minimize errors in size calculations by
classifying the length into groups. Our
aim was to show whether duplications of
SFV were focal or covered a large field of
view of the femoral vein, which is of im-
Figure 3. Coronal venogram of the thigh in three patients illustrates
anomalies of the SFV. (a) Short segment lateral duplication (arrow) of portance to the sonographer. Finally,
the SFV. (b) Long segment lateral duplication (arrow) of the SFV. venography is not necessarily the stan-
(c) Complex venous anatomy of the SFV demonstrates three vessels at dard technique for assessment of detailed
the midthigh level. Arrows indicate duplicated vessels. anatomy, since it depends on adequate
filling of all vessels, which is often not
possible because of technical problems. It
the smaller the vessel, the more difficult strongly correlated with the incidence of also causes the assumption that any ves-
US visualization becomes. venous anomalies in the other limb. This sels that do not fill with contrast material
The finding of multiple vessels within result differs from conclusions in two are not necessarily present.
the popliteal fossa in 44% of patients is of other studies (9,10), both of which were Our results demonstrate that variations
importance for imaging this region. Al- conducted with US. Therefore, the impli- in lower limb venous anatomy are com-
though there is evidence of an increase in cations of this observation are that if du- mon and have important implications for
the incidence of DVT in patients with plications are visualized in one leg, the diagnostic US imaging of suspected DVT.
duplicated SFV, this has not been dem- other leg should be scanned for the pres- The results are in concordance with those
onstrated in previous studies in which ence of these anomalies. of previous venographic studies in demon-
patients with duplicated popliteal veins Little variation was seen in the anat- strating a high degree of variation in lower
were examined (13). At present, we are omy of the calf veins, with the majority limb venous anatomy (7,8). However, the
aware of only several cases highlighting of calf vessels being paired. Greater vari- present study differs with regard to the
the presence of thrombus in one of two ation was seen, however, with respect to population studied and the availability of a
veins in patients with congenital dupli- the position of confluence of veins that large number of bilateral lower limb
cation of the popliteal vein (18,19). Nev- formed the popliteal vein. venograms. The fact that US studies have
ertheless, our study highlights the high Our findings present implications for demonstrated lower frequency rates is a
incidence of multiple vessels but not du- routine imaging of the SFV, a finding not matter of concern and may highlight an
plication in the popliteal fossa. This may shared by all authors. There are advocates important difference in the sensitivity and
increase the likelihood of a missed for the minimal use of US (21–23), which specificity of these two examination tech-
thrombus at US if only a single vein is suggests that US venous studies can be niques. This has considerable implications
visualized. Our findings are similar to safely limited to the popliteal and com- for physician confidence in US as a reliable
those of other authors who reported rates mon femoral veins. However, at least imaging modality for the exclusion of
of 36% (19) and 44% (20) in their series, 20% of proximal thrombi are isolated to DVT. Unless an assessor-blinded study is
with most popliteal vessel duplications the SFV (24 –26), and this finding may be performed to directly compare US and
resulting from the high confluence of the more common when the SFV is dupli- venography with regard to the ability to
posterior tibial and peroneal veins within cated (7). In a review of 269 cases of acute identify venous anomalies, we would rec-
the popliteal space. lower limb proximal DVT, Maki et al (25) ommend to all who perform US for the
Multiple vessels in the popliteal fossa showed that 22% of thrombi were iso- diagnosis of DVT to look for two vessels in
or duplication of the SFV in one limb lated to the SFV, with the remaining 78% the popliteal fossa and for a duplicated

Volume 228 䡠 Number 2 Variations in Lower Limb Venous Anatomy 䡠 447

 SFV; the examination should start at the investigation of deep vein thrombosis. congenital double popliteal vein: case re-
adductor canal and extend to the mid- Radiology 1998; 206:397– 401. port. Angiology 1989; 40:933–936.
9. Gordon AC, Wright I, Pugh ND. Duplica- 19. Rose SC, Zwiebel WJ, Miller FJ. Distribu-
thigh, and the sonographer should look tion of the superficial femoral vein: rec- tion of acute lower extremity deep ve-
both medial and lateral to the main SFV. ognition with duplex ultrasonography. nous thrombosis in symptomatic and
This will prolong an already time-consum- Clin Radiol 1996; 51:622– 624. asymptomatic patients: imaging implica-
Radiology

ing examination, but it will ensure that a 10. Kerr TM, Smith JM, McKenna P, et al. tions. J Ultrasound Med 1994; 13:243–
Venous and arterial anomalies of the 250.
potential venous thrombus is not missed. lower extremities diagnosed by duplex 20. May R. Surgery of the veins of the leg and
scanning. Surg Gynecol Obstet 1992; 175: pelvis. Philadelphia, Pa: Saunders, 1979;
Acknowledgments: We thank Nadine Weiss- 309 –314. 4.
linger and Sophie Combe from Aventis Pharma 11. Samama MM, Cohen AT, Darmon JY, et 21. Lensing AWA, Prandoni P, Brandjes D, et
for providing the venographic data for us to re- al. A comparison of enoxaparin with pla- al. Detection of deep-vein thrombosis by
view. We thank the central reading committee of cebo for the prevention of venous throm- real-time B-mode ultrasonography. N Engl
the MEDENOX study (Chairman, Philippe Gi- boembolism in acutely ill medical pa- J Med 1989; 320:342–345.
rard) for performing the initial evaluation of the tients: prophylaxis in Medical Patients 22. Kim HM, Kuntz KM, Cronan JJ. Optimal
venograms. We thank Qilong Yi for providing with Enoxaparin Study Group. N Engl management strategy for use of compres-
statistical assistance. Travel expenses were pro- J Med 1999; 341:793– 800. sion us for deep venous thrombosis in
vided by Aventis Pharma. 12. Rabinov K, Paulin S. Roentgen diagnosis symptomatic patients: a cost-effective-
of venous thrombosis in the leg. Arch ness analysis. Acad Radiol 2000; 7:67–76.
Surg 1972; 104:134 –144. 23. Cogo A, Lensing AW, Koopman MM, et
References 13. Dona E, Fletcher JP, Hughes TM, Saker K,
al. Compression ultrasonography for di-
1. Lensing AW, Prandoni P, Prins MH, Buller Batiste P, Ramanathan I. Duplicated pop-
agnostic management of patients with
HR. Deep-vein thrombosis. Lancet 1999; liteal and superficial femoral veins: inci-
clinically suspected deep vein thrombo-
353:479–485. dence and potential significance. Aust
sis: prospective cohort study. Br Med J
2. Fraser JD, Anderson DR. Deep venous N Z J Surg 2000; 70:438 – 440.
1998; 316:17–20.
thrombosis: recent advances and optimal 14. Hull RD, Hirsh J, Carter CJ, et al. Pulmo-
24. Frederick MG, Hertzberg BS, Kliewer MA,
investigation with US. Radiology 1999; nary angiography, ventilation lung scan-
211:9 –24. ning, and venography for clinically sus- et al. Can the US examination for lower
3. Kearon C, Julian JA, Newman TE, Gins- pected pulmonary embolism with extremity deep venous thrombosis be ab-
berg JS. Noninvasive diagnosis of deep abnormal perfusion lung scan. Ann In- breviated? A prospective study of 755 ex-
venous thrombosis. Ann Intern Med tern Med 1983; 98:891– 899. aminations. Radiology 1996; 199:45– 47.
1998; 128:663– 677. 15. Girard P, Musset D, Parent F, Maitre S, 25. Maki DD, Kumar N, Nguyen B, Langer JE,
4. Gabella G. Cardiovascular system. In: Wil- Phlippoteau C, Simonneau G. High prev- Miller WT Jr, Gefter WB. Distribution of
liams PL, Bannister LH, et al, eds. Gray’s alence of detectable deep venous throm- thrombi in acute lower extremity deep
anatomy. 38th ed. Edinburgh, Scotland: bosis in patients with acute pulmonary venous thrombosis: implications for
Churchill Livingstone, 1995; 1595. embolism. Chest 1999; 116:903–908. sonography and CT and MR venography.
5. Cockett FB. Abnormalities of the deep 16. May R, Nissl R. The relevance of phlebog- AJR Am J Roentgenol 2000; 175:1299 –
veins of the leg. Post Med J 1954; 30:512– raphy in the diagnostics and treatment of 1301.
522. deep phlebothromboses of the leg. Munch 26. Badgett DK, Comerota MC, Khan MN,
6. Thomas ML. Phelography of the lower Med Wochenschr 1973; 115:2182–2188. Eid IG, Kerr RP, Comerota AJ. Duplex ve-
limb. New York, NY: Churchill Living- [German] nous imaging: role for a comprehensive
stone, 1982; 162–163. 17. Quinn KL, Vandeman FN. Thrombosis of lower extremity examination. Ann Vasc
7. Liu GC, Ferris EJ, Reifsteck JR, Baker ME. a duplicated superficial femoral vein: po- Surg 2000; 14:73–76.
Effect of anatomic variations on deep ve- tential error in compression ultrasound 27. Ascani A, Radicchia S, Parise P, Nenci GG,
nous thrombosis of the lower extremity. diagnosis of lower extremity deep venous Agnelli G. Distribution and occlusiveness
AJR Am J Roentgenol 1986; 146:845– 848. thrombosis. J Ultrasound Med 1990; of thrombi in patients with surveillance
8. Screaton NJ, Gillard JH, Berman LH, 9:235–238. detected deep vein thrombosis after hip
Kemp PM. Duplicated superficial femoral 18. Bronzi G, Venarucci V, Delle Monache G, surgery. Thromb Haemost 1996; 75:239 –
veins: a source of error in the sonographic Scudieri M, Bellagamba G. Thrombosis of 241.

448 䡠 Radiology 䡠 August 2003 Quinlan et al