Journal of Vascular Surgery
Volume 40, Issue 4 , Pages 604-611, October 2004

Re-evaluation of iliac compression syndrome using magnetic resonance imaging in patients with acute deep venous thromboses

  • Douglas G.W. Fraser, MRCP

      Affiliations

    • Department of Cardiology, Queen Elizabeth Hospital, Birmingham, UK
    • Corresponding Author InformationReprint requests: Dr Douglas Fraser, 3 Aboyne Close, Birmingham, UK
    • ,
  • Alan R. Moody, FRCS

      Affiliations

    • Department of Medical Imaging
    • ,
  • Anne Martel, PhD

      Affiliations

    • Department of Medical Physics, Sunnybrook and Women's College Health Sciences Centre, Toronto, Canada
    • ,
  • Paul S. Morgan, PhD

      Affiliations

    • Department of Academic Radiology, University Hospital, Nottingham, UK

Received 5 April 2004; accepted 12 July 2004.

Article Outline

Background

The majority of proximal deep venous thromboses (DVTs) are thought to have propagated as a contiguous column from the calf veins. However, several authors have proposed that ileofemoral DVT commonly originates in the left common iliac vein (LCIV) at a site of compression by the overlying right common iliac artery (RCIA/LCIV compression). This mechanism could explain both the left-sided predominance of ileofemoral DVT and the finding that ileofemoral DVT frequently occurs either in the absence of calf vein thrombosis (isolated ileofemoral DVT) or is not contiguous with calf vein thrombosis (noncontiguous ileofemoral DVT). This mechanism remains unconfirmed.

Objectives

The purpose of this study was to detect RCIA/LCIV compression using multimodal magnetic resonance imaging in thrombosed and patent iliac veins, to determine whether RCIA/LCIV compression occurs more frequently in cases of left ileofemoral DVT than other types of DVT, and to determine if RCIA/LCIV compression is specifically associated with left isolated and noncontiguous ileofemoral DVT.

Patients and methods

This prospective study conducted at the 1355-bed University Hospital included 18 patients with ileofemoral DVT, 23 with femoropopliteal DVT, 15 with isolated calf DVT recruited consecutively, and 28 control patients in whom DVT had been excluded. Interventions included magnetic resonance direct thrombus imaging (MRDTI), venous enhanced peak arterial magnetic resonance venography (VESPA) and magnetic resonance arteriography (MRA) within 48 hours of routine conventional venography (CV). RCIA/LCIV compression of patent LCIVs was assessed using VESPA and MRA; RCIA/LCIV compression of thrombosed LCIVs was assessed using MRDTI and MRA. The extent of calf and popliteal thrombosis was detected using CV; the extent of femoral and iliac thrombosis was detected using VESPA and MRDTI.

Results

RCIA/LCIV compression was more commonly detected in cases of left ileofemoral DVT (9/16 cases) than in cases of left femoropopliteal DVT (1/11 cases; P = .018), right femoropopliteal DVT (2/12 cases; P = .054), left isolated calf DVT (1/9 cases; P = .037), right isolated calf DVT (0/6 cases; P = .046) and control patients (4/28 cases; P = .006). RCIA/LCIV compression was more commonly detected in cases of left isolated ileofemoral DVT (6/6 cases; P = .005), and cases of left noncontiguous ileofemoral DVT (2/2 cases; P = .067) than in cases in which thrombosis was contiguous from the calf to the iliac veins (1/8 cases).

Conclusion

RCIA/LCIV compression was strongly associated with left ileofemoral DVT and was specifically associated with cases that involve independent ileofemoral thrombosis.

 

The majority of lower-limb deep venous thromboses (DVTs) are thought to arise within the calf veins and propagate as a contiguous column into the proximal veins.1, 2, 3, 4, 5 Although femoropopliteal DVT occurs equally in left and right limbs and is contiguous with calf thrombosis in almost all cases, the pattern of ileofemoral thrombosis is frequently inconsistent with this mechanism.4, 6, 7, 8 In one third of cases, ileofemoral DVT occurs either in the absence of calf vein thrombosis (isolated ileofemoral DVT) or is not contiguous with calf vein thrombosis (noncontiguous ileofemoral DVT).4, 6, 7, 8 Ileofemoral DVT is also left sided in 60% to 80% of cases with 85% to 95% of cases of isolated ileofemoral DVT occurring in the left leg.9, 10, 11, 12, 13, 14, 15, 16

These findings could be explained by thrombus formation in the left iliac veins as suggested by May and Thurner,9 and later Cockett et al.11 They suggested that ileofemoral thrombosis commonly originates in the left common iliac vein at a site where the vein is compressed by the right common iliac artery (RCIA/LCIV compression).9, 10, 11, 17 However, this mechanism remains unconfirmed due to difficulties visualizing such compression with conventional techniques in iliac veins that are occluded with fresh thrombosis.18

In this study we have used multimodal magnetic resonance imaging (MRI) to visualize RCIA/LCIV compression in both thrombosed and patent iliac veins. Direct thrombus MRI specifically visualizes acute thrombus and was used to visualize thrombosed iliac veins.19 Venous enhanced peak arterial magnetic resonance venography (VESPA MRV) was used to visualize patent iliac veins and magnetic resonance arteriography (VESPA MRA) was used to visualize the iliac arteries. We hypothesised that RCIA/LCIV compression would be more commonly associated with isolated and noncontiguous ileofemoral DVT than cases of contiguous thrombosis from the calf to the iliac veins (contiguous ileofemoral DVT). The frequency of RCIA/LCIV compression was therefore compared in patients with left ileofemoral DVT and other types of DVT and was compared in patients with contiguous, noncontiguous, and isolated ileofemoral DVT.

Back to Article Outline

Methods 

The study was performed at the University Hospital Nottingham, United Kingdom, where the ethical committee granted approval for the study and all study patients gave written informed consent. Patients were recruited following routine conventional venography (CV) for patients with suspected acute DVT due to the presence of lower-limb symptoms. During the study period CV was used for the diagnosis of DVT in all patients at our institution except for pregnant women and patients with known contrast allergy, who were investigated with venous duplex scanning. The latter patients were therefore excluded from the study. Additional exclusion criteria were the inability to tolerate MRI, contraindications to MRI, and discordant results from MRI and conventional venography.

In the first phase of recruitment (between May 1998 and August 1999) consecutive patients with positive venograms and 1 in 8 patients with negative conventional venograms were recruited. Patients with negative venograms were selected using a random sequence. In the second phase of recruitment (September to December 1999) only consecutive patients with ileofemoral DVT and isolated calf DVT were recruited. This protocol was chosen to equalize the numbers of patients with ileofemoral, femoropopliteal, and isolated calf DVT, and control patients (in whom DVT had been excluded). All patients underwent MRDTI, VESPA MRV, and VESPA MRA within 5 days of conventional venography.

Imaging 

MRDTI is a 3-dimensional spoiled steady-stategradient echo sequence based on the Siemens 3D MP RAGE sequence (TE 4 ms, TR 10.3 ms, 15° flip angle, 1000-ms recovery time between each inversion recovery excitation block).19This sequence is highly T1 weighted and is optimized to display acute thrombus as high signal with a suppressed background. It incorporates a selective water excitation pulse and an inversion time of 20 ms to null blood signal. The scanning time using 2 overlapping, 50-cm imaging blocks and the body coil from the ankle to the inferior vena cava (IVC) was 12 minutes.

VESPA is a 3-dimensional contrast-enhanced gradient echo sequence. Sequential acquisitions before and after a contrast bolus produce background, early arterial phase, and late venous phase images. MRA images are produced by using postprocessing to remove background signal from early arterial phase images. MRV images are produced by using postprocessing to remove arterial and background signal from late venous phase images.20 The sequence is based on the Siemens FISP sequence (fast imaging with steady-state precession; TE 2 ms, TR 5 ms, flip angle 35°). Eight successive acquisitions each lasting 30 seconds were recorded (total imaging time 4 minutes) with a contrast bolus rapidly injected into an antecubital vein halfway through the first acquisition (20 mL of gadopentate dimeglumine, Magnevist, Berlex Laboratories, NJ). Subtraction of the sum of acquisitions 1 and 2 from 7 and 8 was used to produces a selective venogram (VESPA MRV), and subtraction of acquisition 1 from acquisition 2 was used to produce a selective arteriogram (VESPA MRA). In patients with a slow circulation time, acquisition 3 was used instead of acquisition 2. A vacuum extracted beanbag was used to fix the legs in position between acquisitions. Acquisitions were acquired using a 50-cm field of view from the popliteal veins to the IVC with the body coil. A second overlapping field of view from the ankle to proximal superficial femoral vein was used for cases in which full visualization of the proximal veins was not achieved with a single field of view.

CV was performed by cannulating a dorsal pedal vein with a 21-gauge needle and rapidly injecting 50 to 100 mL of iodinated contrast medium (300mg/mL I2) with the patient supine and tipped 30° (feet downward). A tourniquet was applied above the ankle. Anteroposterior and 2 oblique views were obtained of the deep calf and popliteal veins. Views of the femoral and iliac veins were then obtained.

Interpretation of studies 

CV was interpreted by an independent radiologist without knowledge of the other test results. The studies were considered positive if intraluminal filling defects were seen or if persistent nonfilling of veins with a sharp cut-off was detected. CV was used to determine the presence of calf and popliteal thrombosis and to determine if calf thrombosis was contiguous with popliteal thrombosis.

Thrombus was diagnosed by using MRDTI as high venous signal against a suppressed background signal.19 Thrombus was diagnosed by using VESPA MRV in the presence of nonfilling of contrast-enhanced venous segments.20 Both MRDTI and VESPA MRV were used to determine the extent and continuity of iliac and femoral thrombosis and to determine if iliac, femoral, and popliteal thromboses were contiguous.

RCIA/LCIV compression was diagnosed with MRDTI, VESPA MRV and VESPA MRA. Patent left common iliac veins (LCIVs) were visualized by using multiplanar reconstruction of the VESPA MRV datasets, and thrombosed LCIVs were visualized by using multiplanar reconstruction of the MRDTI datasets (Fig 1). The positions of the right common iliac artery (RCIA) and left common iliac artery (LCIA) were visualized in relation to stenoses of the LCIV by overlaying the corresponding VESPA MRA images. RCIA/LCIV compression was defined as a greater than 50% reduction of lumenal cross-sectional area of the LCIV at the point where the LCIV was crossed by the right common iliac artery. A threshold of 50% lumenal narrowing has previously been used for angioplasty of venous spurs thought to be caused by RCIA/LCIV compression after catheter-directed thrombolysis.21

  • View full-size image.
  • Fig 1. 

    Compression of the left common iliac vein by the right common iliac artery (RCIA/LCIV compression) visualized with magnetic resonance venography, arteriography, and direct thrombus imaging in 2 study patients. a, Surface rendered pelvic venogram shows a defect at the origin of the left common iliac vein (arrow). b, Superimposed surface-rendered venogram (blue) and arteriogram (red) show the defect in the LCIV is due to the presence of the right common iliac artery, which crosses and compresses the vein onto the vertebral column at this point (arrow). c, In a second patient, a maximum-intensity projection direct thrombus image of the pelvis shows thrombus filling the left common iliac vein (arrow), left internal iliac vein (double arrow), left external iliac vein (unfilled arrow), and left common femoral vein (unfilled arrow). The black and white arrow shows a cranially directed branch of the internal iliac vein. The dotted lines are the points at which cross-sectional images d and e were constructed. d, Thrombus filling the lumen of the left common iliac vein shows that the vein is flattened (white arrowheads) due to compression by the overriding right common iliac artery (unfilled arrowheads). e, Cross-sectional image showing thrombus filling a more distal portion of the left common iliac vein (arrow) that is not compressed.

Comparisons and statistics 

The frequency of RCIA/LCIV compression was compared in patients with right- and left-sided ileofemoral DVT, right- and left-sided femoropopliteal and isolated calf DVT, and control patients. The frequency of RCIA/LCIV compression was also compared in patients with left isolated ileofemoral DVT, left noncontiguous ileofemoral DVT, and left contiguous ileofemoral DVT.

These analyses were repeated following exclusion of cases associated with risk factors that could have influenced the site of thrombus formation or the pattern of thrombosis. These risk factors included ipsilateral leg trauma or surgery within the previous 3 months, previous ipsilateral DVT, and ipsilateral paresis or immobilization. Hip surgery was considered a potential trigger for both ileofemoral and femoropopliteal DVT, and knee surgery was considered a potential trigger for both calf and femoropopliteal DVT. Specific triggers of pelvic thrombosis also included pregnancy, pelvic malignancy, and pelvic masses.

All significance values were calculated using 2-sided Fisher exact tests.

Back to Article Outline

Results 

Patients with suspected DVT during the study period who were excluded according to the study protocol included 16 patients with suspected DVT associated with pregnancy or known contrast allergy; 58 patients who had either nondiagnostic venography, contraindications to MRI, or claustrophobia; 191 patients with negative venograms; and 12 patients in whom CV diagnosed femoropopliteal DVT in the second phase of recruitment. Further exclusions included 24 patients who refused study entry, 7 patients in whom magnetic resonance scanning had been either nondiagnostic or not tolerated, and 6 patients in whom MRDTI, VESPA MRV, or CV were discordant.

Discordance between CV and MRDTI occurred in 5 patients with possible calf thrombosis. In 3 patients with negative venograms, MRDTI detected thrombus within the gastrocnemius veins. Review of the venograms showed that the gastrocnemius veins had not been visualized and subsequent ultrasound testing also detected gastrocnemius vein thrombosis. In 1 patient with a negative venogram and a subsequent negative ultrasound, MRDTI detected a very small isolated calf thrombus 1cm in length. In 1 patient CV detected isolated calf thrombosis whereas MRDTI was negative; supplementary tests included a negative venous ultrasound and a low D-dimer level in this patient. Visualization of the calf veins using VESPA MRV was not performed in these patients. Discordance between CV, VESPA MRV, and MRDTI occurred in 1 patient in whom CV and MRV detected popliteal vein thrombosis and MRDTI was negative. The patient had an unrecordable D-dimer level (<0.1mg/L), and MRV and CV may have been detecting residual changes from a previous ipsilateral DVT 6 months previously.

The study population included 15 patients with isolated calf DVT (10 left-sided and 6 right-sided), 23 patients with femoropopliteal DVT (11 left-sided and 12 right-sided), 18 patients with ileofemoral DVT (2 right-sided and 16 left-sided) and 28 patients in whom DVT had been excluded (the control group). Forty-four patients were inpatients and 40 patients were outpatients; 76 patients were referred from medical departments and 8 patients were referred from surgical departments. Age ranged from 20 to 87 years and symptom onset varied from 1 to 35 days.

Femoropopliteal DVT was contiguous from the calf to the popliteal and femoral veins in 21 of 23 cases, and contiguous thrombus extended from the short saphenous vein into the popliteal vein in 1 of the 2 remaining cases. There were 2 cases of right contiguous ileofemoral DVT, 8 cases of left contiguous ileofemoral DVT, 6 cases of left isolated ileofemoral DVT, and 2 cases of left noncontiguous ileofemoral DVT. Patients with left isolated ileofemoral DVT were predominantly female (5/6 cases) and their age (mean 41 years) was lower than the age of patients with other types of DVT (mean, 61 years; P = .002).

RCIA/LCIV compression was more commonly detected in cases of left ileofemoral DVT (9/16 cases) than in cases of left femoropopliteal DVT (1/11 cases, P = .018), right femoropopliteal DVT (2/12 cases, P = .054), left isolated calf DVT (1/9 cases, P = .037), right isolated calf DVT (0/6 cases, P = .046) and control patients (4/28 cases, P = .006). The relationship between the pattern of thrombosis and the type of RCIA/LCIV compression is shown in Fig 2. RCIA/LCIV compression was more commonly detected in cases of left isolated ileofemoral DVT (6/6 cases; P = .005) and cases of left noncontiguous ileofemoral DVT (2/2 cases, P = .067) than in cases of left contiguous ileofemoral DVT (1/8 cases). The 2 cases of right contiguous ileofemoral DVT were not associated with RCIA/LCIV compression. RCIA/LCIV compression was detected in a similar proportion of patients with left contiguous ileofemoral DVT (1/8 cases), left or right femoropopliteal DVT (3/23 cases), left or right isolated calf DVT (1/15 cases), and control patients (4/28 cases).

  • View full-size image.
  • Fig 2. 

    Summary of results showing the number of patients with and without compression of the left common iliac vein (LCIV) by the right common iliac artery (RCIA) according to the extent and pattern of thrombosis. Type 1 was compression of the origin of the LCIV by the RCIA. Type 2 was compression of the origin of the LCIV by the RCIA together with compression of the midpoint of the LCIV by the left common iliac artery. contiguous fempop, Contiguous thrombus from the calf to the femoral veins; noncontig fempop, femoropopliteal and calf thrombus that are not contiguous; contiguous ileofem, contiguous thrombus from the calf to the iliac veins; noncontiguous ileofemoral, iliac/ileofemoral thrombus and distal thrombosis that are not contiguous.

Cases of DVT associated with other risk factors that could have influenced the site or pattern of thrombosis included 2 cases of left isolated ileofemoral DVT (both postpartum), 4 cases of left contiguous ileofemoral DVT (pelvic or hip surgery, 3 cases; pelvic malignancy, 1 case), 2 cases of left femoropopliteal DVT (left hemiplegia, 1 case; prior left infrainguinal DVT, 1 case), 1 case of left isolated calf DVT (prior left leg DVT), 2 cases of right femoropopliteal DVT (lower-leg surgery, 1 case; prior right-leg DVT, 1 case), and 2 cases of right isolated calf DVT (ankle surgery, 1 case; knee surgery, 1 case). Following exclusion of these cases, RCIA/LCIV compression remained significantly more commonly associated with left ileofemoral DVT (6/10 cases) than with left femoropopliteal or isolated calf DVT (1/18 cases, P = .003), right femoropopliteal or isolated calf DVT (2/14 cases, P = .032), and control patients (4/28 cases, P = .010). RCIA/LCIV compression also remained more commonly associated with left isolated ileofemoral DVT (4/4 cases, P = .029) and left noncontiguous ileofemoral DVT (2/2 cases, P = .067) than left contiguous ileofemoral DVT (0/4 cases).

Thrombus extended into the left or right common iliac veins in 6 of 10 cases of contiguous ileofemoral DVT. Thrombus extended into the LCIV and up to the site of RCIA/LCIV compression in 6 of 6 cases of left isolated ileofemoral DVT and 2 of 2 cases of left noncontiguous ileofemoral DVT. Distal extension of thrombus initially located in the LCIV was directly observed in 2 study patients (Fig 3). One patient had isolated LCIV thrombosis demonstrated on routine venography that was erroneously reported as negative and entered the study only after a second venogram, which was requested due to worsening symptoms. This showed thrombus extension to the common femoral vein; no thrombus was detected in the distal veins. Disparate thrombus in the LCIV and deep calf veins was present at study entry in a second patient, who was rescanned due to worsening symptoms 5 days later and was found to have thrombus extending from the LCIV to the common femoral vein. The extent of calf thrombosis had not changed.

  • View full-size image.
  • Fig 3. 

    Propagation of thrombus distally from the left common iliac vein in 2 study patients. a, Venogram at presentation shows thrombus filling of the left external iliac vein and retrograde filling of the left internal iliac vein (unfilled arrow). There is no filling of the left common iliac vein (LCIV) with a sharp cut-off due to isolated LCIV thrombus (arrow). b, Repeat venography 2 days later shows that thrombus has extended distally from the LCIV to fill both the left external iliac and common femoral veins (arrows). c, Magnetic resonance pelvic venogram at presentation in a second patient shows a filling defect in the LCIV extending into the inferior vena cava due to thrombus (arrows). Calf thrombus was also present (not shown). d, Repeat magnetic resonance venogram 5 days later shows a filling defect extending from the inferior vena cava distally to the common femoral vein due to distal extension of thrombus into the left external iliac and common femoral veins (arrows).

Two types of RCIA/LCIV compression were observed. Type 1 was compression of the origin of the LCIV by the RCIA. Type 2 was compression of the origin of the LCIV by the RCIA together with compression of the midpoint of the LCIV by the LCIA. Frequently, the LCIV was collapsed between these 2 sites of compression. Type 2 compression approximated to Cockett et al's description of LCIV compression by the aortic bifurcation.11 The mean percentage stenosis in cases of RCIA/LCIV compression was 76%, with a standard deviation of 10%. In 2 patients in whom DVT had been excluded, LCIV stenoses <50% (24% and 22% respectively) were detected at the site of crossing by the right common iliac artery.

Thrombophilia screening that included activated protein C (APC) resistance,prothrombin 20210A gene mutation, antithrombin deficiency, protein C and S levels, and lupus anticoagulant was performed in 29 patients. Abnormal results were detected in a similar proportion of patients (2/8 cases) with isolated or noncontiguous ileofemoral DVT (factor V Leiden and prothrombin 20210A gene mutation respectively) and other types of DVT (5/21 cases).

Back to Article Outline

Discussion 

RCIA/LCIV compression was more commonly detected in patients with left ileofemoral DVT. As we had hypothesized, RCIA/LCIV compression was specifically associated with left isolated and left noncontiguous ileofemoral DVT and was present in all of such cases. These relationships remained following exclusion of cases with other potential localizing causes for thrombosis and were unrelated to the results of thrombophilia testing. The pattern of thrombosis in the latter patients was consistent with thrombus formation at the site of compression. These results support and extend the findings of May and Thurner9 and Cockett et al11 by suggesting that RCIA/LCIV specifically leads to ileofemoral DVT that involves independent ileofemoral thrombosis.

This is the first prospective study to relate consecutive cases of ileofemoral thrombosis to the presence of RCIA/LCIV compression. This is also the first study to relate the pattern of ileofemoral thrombosis to the presence of RCIA/LCIV compression. The extended recruitment of patients with ileofemoral and calf DVT and the exclusion of the majority of patients in whom DVT had been excluded were intended to equalize the numbers of patients with isolated calf, femoropopliteal, and ileofemoral DVT and control patients. Control patients were selected by a random sequence, and cases of ileofemoral, femoropopliteal and isolated calf DVT were consecutive. However, due to the small study size, analysis of subgroups was limited and there were only 2 cases of right ileofemoral DVT and only 2 cases of left noncontiguous DVT. The association between RCIA/LCIV compression and noncontiguous ileofemoral DVT will therefore require further confirmation.11 The extent and continuity of calf and popliteal thrombosis was assessed by using CV, which remains the criterion standard test at this level. The extent and continuity of femoral and iliac thrombosis were assessed by using MRDTI and VESPA rather than venography because venography frequently does not display the proximal extent of femoral and iliac thrombosis and is frequently of poor quality within the pelvis.20, 22, 23 Several authors have considered MRI techniques as the reference standard for the evaluation of pelvic thrombosis.24, 25, 26, 27 We have shown that both MRDTI and VESPA MRV are highly accurate diagnostic tests and display greater proximal extension of thrombosis than conventional venography.19, 20, 23 In a study of 101 patients with acute thrombosis, MRDTI detected femoral and iliac thrombosis with sensitivities and specificities of 97% to 100% and κ values for interobserver variability of 0.96 and 0.98 respectively; accuracy was also maintained below knee.23 In a study of 55 patients, femoral and iliac vein thrombosis was detected by using VESPA MRV with sensitivities and specificities of 97% to 100% and κ values for interobserver variability of 0.85 and 0.97 respectively.20 However, the inability to blind the analysis with respect to the extent of thrombosis and the use of different MRI sequences for the assessment of patent and thrombosed LCIVs are potential sources of bias. To minimize this bias we used 2 reviewers and a quantitative rather than qualitative definition of RCIA/LCIV compression.

The continuity of calf thrombosis with femoropopliteal and ileofemoral thrombosis in this study was similar to the findings of previous studies. Femoropopliteal DVT was contiguous with calf thrombosis in 21 of 23 cases (91%), and contiguous, non-contiguous, and isolated ileofemoral thrombosis were found in 10 of 18 (56%), 2 of 18 (11%) and 6 of 18 (32%) patients with ileofemoral DVT, respectively. Previous venographic studies have found that femoropopliteal thrombosis was contiguous with calf thrombus 88% to 100% of cases, and contiguous, non-contiguous, and isolated ileofemoral thrombosis occurred in 54% to 78%, 3% to 24% and 17% to 25%, respectively, of patients with ileofemoral DVT.4, 6, 7, 8

Thrombus extended up to the site of compression in all of the cases left-sided isolated and noncontiguous ileofemoral DVT. In 1 case of isolated ileofemoral DVT and 1 case of noncontiguous ileofemoral DVT, thrombus extension distally from the left common iliac vein into the left external iliac and proximal femoral veins was directly observed (Fig 3). These findings are consistent with the proposal by Cockett et al11 that thrombus formation occurs at the site of RCIA/LCIV compression in the left common iliac vein and extends distally to the femoral veins. Cockett and colleagues11, 17 investigated patients with severe postphlebitic disease and a history of ileofemoral DVT by using iliac phlebography, and concluded that RCIA/LCIV compression was not only the cause of the iliac thrombosis but also prevented its recanalization. Although these findings have remained unconfirmed, defects attributed to RCIA/LCIV compression have been detected in up to 60% of cases of ileofemoral DVT following thrombectomy and catheter-directed thrombolysis and are successfully treated with angioplasty and stenting.16, 21, 28, 29, 30

It has been suggested that RCIA/LCIV compression leads to thrombosis within the left common iliac vein either by trapping small emboli from thrombus in the distal veins or by de novo thrombus formation at the site of compression.11, 31Trapping of emboli from distal thrombosis may explain the occurrence of disparate thrombosis in the calf and ileofemoral veins in cases of noncontiguous ileofemoral DVT. The mechanism of de novo thrombus formation at the site of compression is unknown, but may be related to turbulence and fibrosis of the venous wall at this site.11, 32 A corollary of this situation is found in the retina; many cases of retinal vein thrombosis are believed to be caused by compression of the retinal veins by overlying retinal arteries at sites of arteriovenous crossover.33, 34 Venous endothelial damage secondary to turbulent flow has been demonstrated at these sites in the retina and is thought to act as the trigger for thrombosis.35 Endothelial lesions in the left common iliac vein associated with RCIA/LCIV, such as intraluminal webs and spurs reported by Cockett et al11 and others32 were not clearly evident in this study but may have contributed to the reduction of lumenal area recorded.

The occurrence of RCIA/LCIV compression in young women has been described previously and has been related to changes that occur during maturation of the pelvis during puberty.36 Accentuation of the lumbar lordosis pushes the lower lumbar vertebrae forward and compresses the overlying LCIV against the RCIA, which crosses the vein at this point. These changes are further accentuated during pregnancy. This mechanism could explain the high proportion of left-sided isolated ileofemoral DVT found in young women in this study and in previous studies of DVT associated with pregnancy and the oral contraceptive pill.12, 13, 14, 15

Back to Article Outline

Conclusion 

We were able to detect RCIA/LCIV compression in both thrombosed and patent iliac veins by using MRDTI and VESPA imaging sequences. RCIA/LCIV compression was specifically associated with left noncontiguous and left isolated ileofemoral DVT. The pattern of thrombosis in these patients suggested that thrombus formation had occurred at the site of RCIA/LCIV compression. These results support the proposal that RCIA/LCIV compression is a major cause of ileofemoral DVT and suggest that the presence of RCIA/LCIV compression can be predicted if either isolated or noncontiguous ileofemoral DVT is present.

Back to Article Outline

References 

  1. Fraser JD , Anderson DR . Deep venous thrombosis: recent advances and optimal investigation with US . Radiology . 1999;211:9–24
  2. Hirsh J , Hoak J . Management of deep vein thrombosis and pulmonary embolism (A statement for healthcare professionals. Council on Thrombosis (in consultation with the Council on Cardiovascular Radiology), American Heart Association) . Circulation . 1996;93:2212–2245
  3. Kearon C , Julian JA , Newman TE , Ginsberg JS . Noninvasive diagnosis of deep venous thrombosis (McMaster Diagnostic Imaging Practice Guidelines Initiative) . Ann Intern Med . 1998;128:663–677
  4. Cogo A , Lensing AW , Prandoni P , Hirsh J . Distribution of thrombosis in patients with symptomatic deep vein thrombosis (Implications for simplifying the diagnostic process with compression ultrasound) . Arch Intern Med . 1993;153:2777–2780
  5. Gibbs NM . Venous thrombosis of the lower limbs with particular reference to bed rest . Br J Surg . 1957;45:215–236
  6. Nylander G , Olivecrona H . The phlebographic pattern of acute leg thrombosis within a defined urban population . Acta Chir Scand . 1976;142:505–511
  7. Stamatakis JD , Kakkar VV , Lawrence D , Bentley PG . The origin of thrombi in the deep veins of the lower limb: a venographic study . Br J Surg . 1978;65:449–451
  8. Nicolaides AN , Kakkar VV , Field ES , Renney JT . The origin of deep vein thrombosis: a venographic study . Br J Radiol . 1971;44:653–663
  9. May R , Thurner J . The cause of the predominantly sinistral occurrence of thrombosis of the pelvic veins . Angiol . 1957;8:419–427
  10. Mavor GE , Galloway JM . Iliofemoral venous thrombosis (Pathological considerations and surgical management) . Br J Surg . 1969;56:45–59
  11. Cockett FB , Thomas ML , Negus D . Iliac vein compression (Its relation to iliofemoral thrombosis and the post-thrombotic syndrome) . Br Med J . 1967;2:14–19
  12. Kierkegaard A . Side and site of deep vein thrombosis in women using oral contraceptives . Acta Obstet Gynecol Scand . 1985;64:399–402
  13. Bergqvist A , Bergqvist D , Hedner U . Pregnancy thrombosis—clinical manifestations . Lakartidningen . 1982;79:901–902
  14. Bergqvist A , Bergqvist D , Hallbook T . Deep vein thrombosis during pregnancy (A prospective study) . Acta Obstet Gynecol Scand . 1983;62:443–448
  15. Bergqvist D , Hedner U . Pregnancy and venous thrombo-embolism . Acta Obstet Gynecol Scand . 1983;62:449–453
  16. Mewissen MW , Seabrook GR , Meissner MH , Cynamon J , Labropoulos N , Haughton SH . Catheter-directed thrombolysis for lower extremity deep venous thrombosis: report of a national multicenter registry . Radiology . 1999;211:39–49
  17. Cockett FB , Thomas ML . The iliac compression syndrome . Br J Surg . 1965;52:816–821
  18. Verhaeghe R . Iliac vein compression as an anatomical cause of thrombophilia: Cockett's syndrome revisited . Thromb Haemost . 1995;74:1398–1401
  19. Moody AR , Pollock JG , O'Connor AR , Bagnall M . Lower-limb deep venous thrombosis: direct MR imaging of the thrombus . Radiology . 1998;209:349–355
  20. Fraser DG , Moody AR , Davidson IR , Martel AL , Morgan PS . Deep venous thrombosis: diagnosis by using venous enhanced subtracted peak arterial MR venography versus conventional venography . Radiology . 2003;226:812–820
  21. Semba CP , Dake MD . Iliofemoral deep venous thrombosis: aggressive therapy with catheter-directed thrombolysis . Radiology . 1994;191:487–494
  22. Coel MN . Adequacy of lower limb venous opacification: comparison of supine and upright phlebography . AJR Am J Roentgenol . 1980;134:163–165
  23. Fraser DG , Moody AR , Morgan PS , Martel AL , Davidson I . Diagnosis of lower-limb deep venous thrombosis: a prospective blinded study of magnetic resonance direct thrombus imaging . Ann Intern Med . 2002;136:89–98
  24. Spritzer CE , Arata MA , Freed KS . Isolated pelvic deep venous thrombosis: relative frequency as detected with MR imaging . Radiology . 2001;219:521–525
  25. Carpenter JP , Holland GA , Baum RA , Owen RS , Carpenter JT , Cope C . Magnetic resonance venography for the detection of deep venous thrombosis: comparison with contrast venography and duplex Doppler ultrasonography . J Vasc Surg . 1993;18:734–741
  26. Erdman WA , Jayson HT , Redman HC , Miller GL , Parkey RW , Peshock RW . Deep venous thrombosis of extremities: role of MR imaging in the diagnosis . Radiology . 1990;174:425–431
  27. Redman HC . Deep venous thrombosis: is contrast venography still the diagnostic “gold standard”? . Radiology . 1988;168:277–278
  28. Juhan CM , Alimi YS , Barthelemy PJ , Fabre DF , Riviere CS . Late results of iliofemoral venous thrombectomy . J Vasc Surg . 1997;25:417–422
  29. Mickley V , Schwagierek R , Rilinger N , Gorich J , Sunder-Plassmann L . Left iliac venous thrombosis caused by venous spur: treatment with thrombectomy and stent implantation . J Vasc Surg . 1998;28:492–497
  30. AbuRahma AF , Perkins SE , Wulu JT , Ng HK . Iliofemoral deep vein thrombosis: conventional therapy versus lysis and percutaneous transluminal angioplasty and stenting . Ann Surg . 2001;233:752–760
  31. Jaszczak P , Mathiesen FR . The iliac compression syndrome . Acta Chir Scand . 1978;144:133–136
  32. Negus D , Fletcher EW , Cockett FB , Thomas ML . Compression and band formation at the mouth of the left common iliac vein . Br J Surg . 1968;55:369–374
  33. Frangieh GT , Green WR , Barraquer-Somers E , Finkelstein D . Histopathologic study of nine branch retinal vein occlusions . Arch Ophthalmol . 1982;100:1132–1140
  34. Weinberg D , Dodwell DG , Fern SA . Anatomy of arteriovenous crossings in branch retinal vein occlusion . Am J Ophthalmol . 1990;109:298–302
  35. Clemett RS . Retinal branch vein occlusion (Changes at the site of obstruction) . Br J Ophthalmol . 1974;58:548–554
  36. Calnan J , Kountz S , Pentecost B , Shillingford J , Steiner R . Venous obstruction in the aetiology of lymphoedema praecox . Br Med J . 1964;2:221–226

 Supported by the British Heart Foundation.

 Competion of interest: none.

PII: S0741-5214(04)00958-9

doi:10.1016/j.jvs.2004.07.039

Journal of Vascular Surgery
Volume 40, Issue 4 , Pages 604-611, October 2004

Article Tools

* Image Usage & Resolution