Venous Ultrasound for DVT

 Abstract

Introduction:  Practices vary widely among sonography laboratories in the imaging and evaluation of calf thrombus, contributing to diagnostic and treatment uncertainty.  Our goal was to isolate and determine patterns of deep vein thrombus to question the level of sonography required for best diagnosis.

Materials and Methods:  We retrospectively evaluated all patients that had a positive venous duplex exam from 2004 to 2013. Our lab performed 11,503 venous studies during the 8 year period, with 382 showing, via symptom review and reporting, acute first-time thrombus. 

Results: Mean age 62 years 8 months old, 209 Females (54.71%), 171 male (44.76%), 221 left sided (57.85%), 161 right sided (42.15%).  Percentages for the following segments: Above the knee (139, 36.39%), Calf (375, 98.17%) Superficial (54, 14.14%), IVC (14, 3.66%) Iliac (36, 9.42%), CFV (67, 17.54%), PFV (7, 1.83%), FV (114, 29.84%), Popliteal AK (123, 32.20%), Popliteal Fossa (142, 37.17%), Popliteal BK (148, 38.74%), Gastrocnemius Veins (141, 36.91%), Peroneal (243, 63.61%), Posterior Tibial Vein (191, 50.00%) ,Soleal (109, 28.53%), Anterior Tibial Veins (3, 0.79%), Great Saphenous Vein (38, (10%), Small Saphenous Vein (18, 5%) and Varicosities (8, 2%).

Conclusion:  Calf vein thrombus is found with above the knee thrombus (98.17%). Peroneal (63.61%), Posterior Tibial (50%), Gastrocnemius (36.91%) and Soleal (28.53%) veins should be routinely imaged. Binomial logistic regression was conducted with eleven predictor variables significantly predicted proximal thrombus (above the knee).  A test with all eleven predictor variables, compared with the null model, was significant (x² (14) = 288.511, p < .001), showing that these variables predict above-the-knee thrombus better than without them (or by chance). We have developed a tool that shows the potential trajectories or tracks of a given patient's DVT (with associated probabilities), based on aggregated patient data. 

Introduction:

Collectively deep vein thrombosis (DVT) and pulmonary embolism (PE) are known as venous thromboembolism (VTE) [1]. Although the exact incidence of VTE is unknown, it is believed that there are approximately 1 million cases in the United States each year, many of which represent recurrent disease [2]. It is estimated that anywhere from 200,000 up to 600,000 Americans will suffer from deep vein thrombosis (DVT) and pulmonary embolism (PE). Additionally, 300,000 of these patients die. [3]   In the United States more people die each year from PE than motor vehicle accidents, breast cancer and AIDS combined [3, 4, 5, 6 and 7].

Practices vary widely among sonography laboratories in the imaging and evaluation of calf thrombus, contributing to diagnostic and treatment uncertainty.  Our goal was to determine the distribution of deep vein thrombus in the lower extremity to assist in answering the following question. What level of sonography is required for best practice and diagnosis when evaluating the lower extremity for deep vein thrombosis?

 It has been our collective experience performing and interpreting venous color flow duplex ultrasound that the location of thrombus presents primarily in the calf and progresses up the lower extremity. Initially in our vascular careers, evaluation of tibial and peroneal vein was performed, however the muscular calf veins (soleal and gastrocnemius) were not included in the routine protocol.  Approximately 12 years ago, the muscular calf veins (MCV) were added to the venous protocol for imaging of the lower extremity for deep vein thrombosis and in the evaluation for venous insufficiency. The incidence of thrombus identified after making the change was significant. The realization that we had missed or potentially missed a significant amount of thrombus before making the change in protocol was of great concern and the impetus for this study.

Providing venous training across the United States, we have witnessed wide variations between facilities, service lines and protocols in the evaluation of the lower extremity for DVT. We have found that in environments in which calf vein thrombus is not deemed important, scanning techniques to evaluate the calf veins often have not been developed and/or taught. In facilities that calf vein thrombus is considered significant the skill of the sonographer has been developed and the testing is very accurate and reliable. It became apparent that the discrepancy in testing leads to subsequent treatment and in many cases the lack of treatment in regard to calf vein thrombus. Our goal was to quantify the amount of thrombus found in each named vein of the lower extremity and to determine the relationship between calf vein thrombus and above the knee thrombus if any. Conflicting clinical results have resulted in differing opinions on the need to test for calf deep vein thrombus (CDVT) as well as treat it. During your review of the literature we determined that there are three prevailing views on how diagnostic imaging of the lower leg is and should be performed. 

Review of Literature: 

The first view is that compression ultrasound (CUS) is sensitive for proximal DVT and that proximal DVT has traditionally been associated with a higher risk for PE than CDVT [8]. From this perspective, imaging of the calf veins is of limited importance as DVT below the knee often resolves spontaneously and rarely is associated with pulmonary embolism or adverse outcomes [8]. (ACR–AIUM–SRU Practice Guideline for the Performance of Peripheral Venous Ultrasound Examination Res. 31 – 2010) suggests that for normal examinations, grey scale images should be recorded without and with compression of the common femoral, saphenous-femoral junction, femoral and popliteal veins, and Doppler of the common femoral and popliteal veins.  Symptomatic areas such as the calf generally require additional evaluation and additional imaging; however, this is left up to the user and/or organization to define [8, 9]. The ACR appropriateness Guide states “DVT that is limited to the infra-popliteal calf veins (i.e., below-the-knee or distal DVT) often resolves spontaneously and is rarely associated with pulmonary embolism or other adverse outcomes” [10, 11, 12, 13]. Investigators in the (Clinical Outcomes of Untreated Symptomatic Patients with Negative Findings on Sonography of the Thigh for Deep Vein Thrombosis: Our Experience and a Review of Literature) stated the following “We find that most evaluation of the calf to exclude a DVT have indeterminate findings because the deep veins of the calf cannot be completely assessed due to their small size relative to thigh veins” [9].  The article goes on to say that there are high rates of indeterminate studies for DVT has been reported [9, 14, 15]. The authors of the study found that only 15% of their calf evaluations were diagnostic and (84.2%) were considered indeterminate. The two studies referenced in regard to imaging of the calf veins were published in 1990 [15] and 1995 [16] respectively.  Both of these finding support the long held view that DVT below the knee is not significant or important.

Upon close examination of the studies used to come to this conclusion (Clinical Outcomes of Untreated Symptomatic Patients with Negative findings on Sonographer of the thigh for Deep Vein Thrombosis: Our Experience and a Review of the Literature) shows that the investigators found that most evaluations of the calf to exclude DVT have indeterminate findings because of the small size.  The study also used references from studies from 1989 and 1995 both authored by JF Polak [16, 17] when saying that calf vein examination are 84% indeterminate [9]. Polak in 1996 came back and published and stated that recent studies showed estimated sensitivity of 93% [18]. He further states that almost a third of patients with calf vein deep vein thrombosis have a recurrence of thromboembolic events when not treated with Coumadin [18, 19].  Although Gottlieb et al [9] published his finding in 1999 and referenced Polak’s work; he did not include the most recent data at that time, which would have contradicted his original findings.  As we looked further into the practice guidelines it became apparent that references are either of older guidelines or of studies that also did not routinely image the calf veins or had poor success imaging calf veins.  The ACR guideline do recognize that 80% of PE’s are from deep vein thrombosis of the lower leg.  Interestingly Gottlieb et al [9]study also showed a very low rate of PE (0.7%) compared to studies of isolated calf vein thrombus that showed a PE rate of 7.9% for isolated infra-popliteal DVT’s alone [20]. The rate of PE was 11.4% for femoral-popliteal deep vein thrombus in the same study and a PE rate of 18.1% in multilevel DVT’s involving both segments [20]. The study by Alhalbouni et al [20] published in 2011, included 4035 patients, of which 3146 were of hospital patients. They concluded that there was no statistically difference (p=0.27) in the risk for PE between isolated femoral-popliteal and isolated infra-popliteal DVT. It was noted that a significant number of those with isolated muscular calf veins (soleal and gastrocnemius) developed PE. They concluded that scans should include infra-popliteal veins to include the soleal and gastrocnemius. [20] The Gottlieb et al study only involved 120 direct patients and only 13 patients underwent a second examination because of persistent symptoms. Due to the small sample size the study then pulled from four other literary studies involving 1797 patients; however three of the studies did not include dedicated examination of the calf. Only 12.5% of the patients in the combined analysis (15 out of 120) had a DVT upon reexamination due to high clinical suspicion of DVT. The three studies were performed in 1990 Vaccari et al [21], in 1989 Sarpa et al [22] and in 1993 Heijboer et al [23].  Gottlieb et al [9] did acknowledge that there are three prevailing views in the imaging community, one that demands a high level of accuracy in evaluation the calf and others who accept follow up ultrasound or venography after calf exams with indeterminate findings in patients with persistent symptoms [9]. He also acknowledges that some physicians do not consider calf thrombus to have any risk of embolization to the lungs and do not require any further evaluation after the initial sonogram. Gottlieb also acknowledge that differing rates of pulmonary emboli could result from the various approaches to surveying the calves for DVT and suggest that a uniform approach, to the examination of the symptomatic patient with suspected lower extremity DVT, would be valuable in reducing practice variation [9].

The second view emerging from the literature is that the imaging of the calf veins (posterior tibial and peroneal veins) should be routine and that the soleal and gastrocnemius veins (known as the muscular calf veins, or MCV) are to be imaged when indicated or if included in the facilities protocol. This view is reflected in the current (IAC Standards and Guidelines for Vascular Accreditation Section 4B: Peripheral Venous Testing). [25] The guideline requires imaging of the Common femoral, femoral (proximal, mid and distal), popliteal, posterior tibial, and peroneal veins. The guideline further states that imaging of the (MCV) may be required by laboratory protocol or when indicated along with the common iliac, external iliac, great saphenous, small saphenous, proximal deep femoral, anterior tibial, perforating veins, and inferior vena cava. [25] While this is the predominate protocol used in dedicated vascular laboratories, it allows for imaging of the muscular calf veins based on the choice of the facility and/or end user. This is the primary protocol that was used by the researchers prior to adding routine imaging of the muscular calf veins. It is our experience that many labs with this protocol often have intra-department variation in regard to imaging of the gastrocnemius and soleal veins.  

The third prevailing view requires imaging of all the lower extremity veins to include the deep calf and muscular calf veins (PTV, Peroneal, Gastrocnemius and Soleal). [4, 5, 6, 8, 10]  Espousing this view are mostly vascular laboratories that are ICAVL accredited or adhere to their standards at minimum but have identified the need to image the muscular calf veins. [24] Those studies which included imaging of all of the calf veins showed an appreciative value and increase in distal DVT when compared studies that did not perform imaging of the deep and muscular calf veins.  It is estimated that approximately 40% of the patients with acute isolated calf DVT would be judged to have normal color flow duplex scan (CFDS) examinations results if muscular veins of the calf were not imaged. [27] In studies where MCV alone were imaged, distal DVT was greatly appreciated over proximal DVT [27]. Studies, in which laboratories looked at all the deep and muscular calf veins, show that DVT is not only present at a higher rate but that the thrombus also propagates at a higher rate [28, 29, 30, 20, 24, 31, 26, 32, 27, and 33]. The frequency of distal involvement greatly exceeds that of a proximal involvement in patients with DVT. [28]

Materials and Methods:

 We retrospectively evaluated all patients that had a positive venous duplex exam from 2004 to 2013. Our lab performed 11,503 venous studies during the 8 year period, with 382 showing, via symptom review and reporting, acute first-time thrombus. 

Our lab performed 11,503 venous studies during the 8 year period, with 3,469 positive exams (for DVT and/or venous insufficiency), the studies included 1,227 patients, with 382 showing, via symptom review and reporting, acute and first-time thrombus. For the purpose of this study, patients with chronic deep vein thrombosis or indeterminate thrombus were excluded, as recanalization in such patients may occur in different vein segments at differing rates. 

All of the patients were tested in an ambulatory outpatient setting and were symptomatic at the time of evaluation; no serial monitoring for non-symptomatic patients was performed.  Scanning included color-flow duplex scanning with compression ultrasound used as the primary indicator of thrombus. Imaging was performed on of all of the named veins of the lower extremity including the common femoral, femoral, popliteal, gastrocnemius, peroneal, posterior tibial, soleal, and when indicated, on the anterior tibial, iliac veins (common and external), and IVC.  All tests were performed by experienced Registered Vascular Technologists within a dedicated vascular lab setting.  Our reporting segmented the popliteal vein into three categories: popliteal above the knee, popliteal fossa, and popliteal below the knee.  For analysis and continuity with other studies that only studied above the knee DVT, we consider the above knee popliteal vein as the indication of thigh DVT.

Results

During the retrospective analysis of the acute DVT in our study we found the following results. Studies those were positive for ACUTE deep vein thrombosis. Total of 1227 studies from 2004-2013 were surveyed with a total of 382 fitting the description.   The mean age 62 years 8 months old, 209 Females (54.71%), 171 male (44.76%), 221 left sided (57.85%), 161 right sided (42.15%).  Percentages for the following segments: Above the knee (139, 36.39%), Calf (375, 98.17%) Superficial (54, 14.14%), IVC (14, 3.66%) Iliac (36, 9.42%) CFV (67, 17.54%) PFV (7, 1.83%) FV (114, 29.84%) Popliteal AK (123, 32.20%), Popliteal Fossa (142, 37.17%) Popliteal BK (148, 38.74%) Gastrocnemius Veins (141, 36.91%) Peroneal (243, 63.61%) Posterior Tibial Vein (191, 50.00%) Soleal (109, 28.53%) Anterior Tibial Veins (3, 0.79%) Great Saphenous Vein (38, (10%) Small Saphenous Vein (18, 5%) and Varicosities (8, 2%). The analysis showed a strong association with the calf (98.17%) of all cases and thigh DVT was only seen without calf involvement 1.83% of the time. It was also apparent during the collection of data that visual patterns of thrombus were present.

Binomial logistic regression was conducted to assess whether the eleven predictor variables significantly predicted proximal thrombus (above the knee). Three predictors were derived from demographic variables: age (with four groups: 10-49, 50-64, 65-79, 80+), sex (male/female), and which leg the thrombus was observed in (left/right). Eight predictor variables were taken from observation of distal thrombus (below the knee) in the popliteal fossa (POP FOSSA), popliteal below knee (POP BK), gastrocnemius vein (GASTROC), peroneal vein (PERO), posterior tibial vein (PTV), soleal vein (SOLEAL), anterior tibial vein (ATV), and small saphenous vein (SSV). Data from 382 cases were included in the analysis.

A test with all eleven predictor variables, compared with the null model, was significant (x² (14) = 288.511, p < .001), showing that these variables predict above-the-knee thrombus better than without them (or by chance). Moreover, the model correctly predicted 90.8% of those cases where proximal thrombus had not been observed, and 89.5% of those cases where it had been; overall, the model predicted 90.3% of observed cases.

Table 1 summarizes the raw binary logistic regression coefficients, as well as Wald statistics and odds ratios (Exp (B)) for each of the eleven predictor variables. The table shows that POP FOSSA thrombus is a statistically significant predictor of above-the-knee thrombus (Exp (B) = 0.029, x² (1) = 37.469, p < .001).  Similarly, POP BK thrombus is a statistically significant predictor of above-the-knee thrombus, as well (Exp (B) = 0.128, x² (1) = 6.844, p = .009). Additionally, the LEG variable was a good predictor of above-the-knee thrombus ((Exp (B) = 2.517, x² (1) = 5.763, p = .016). Interestingly, SOLEIL thrombus is approaching statistical significance ((Exp (B) = 2.278, x² (1) = 3.399, p=.065)). All of the other Wald chi-square tests were not significant. 

TABLE 1

Binomial Logistic Regressing Predicting Proximal Thrombus

Discussion

Currently in the United States the American College of Radiology does not require that calf veins be imaged, that symptomatic areas such as the calf generally require additional evaluation and additional imaging. This is left up to the user and/or organization to define [13].  In our data base if we had not selected to image the calf veins we would have missed 54% of the identified thrombus. Our findings statistically validate the position that deep vein thrombosis originates in the calf most of the time (98.17%).

Our findings showed that thrombus found in the popliteal fossa is a statistically significant predictor of above-the-knee thrombus (Esp. (B) = 0.029, x2 (1) = 37.469, p < .001). This supports the use of the two point DVT check as a screening tool in emergency medicine [35]. The data also shows that relying solely on the two point DVT check (common femoral vein and popliteal fossa) would have resulted in 63% to have been initially interpreted as normal or negative. Earlier studies involving calf imaging were performed at a time when it was not a requirement for a sonographer to hold a vascular credential. Medicare requires a vascular credential for reimbursement in many states and the Intersocietal Accreditation Commission (IAC) will require all technical staff to hold a vascular credential by 2017.

In facilities that do not currently routinely image calf veins, the sonographer in this setting may not have had the training and/or experience imaging calf veins. The development of this skill, while technical, requires minimal additional training and expense. Increased accuracy of the venous exam should coincide with decreased use of venography, CT venography and less negative surveillance ultrasounds of the thigh, thus lowering the total cost of services. The early detection of isolated or multifocal calf DVT and early treatment may decrease the amount of post phlebitic changes in the limb and cut the overall cost associated with chronic venous insufficiency.  Studies have shown that almost a third of patients with calf vein deep vein thrombosis have a recurrence of thromboembolic events when not treated with Coumadin (Du Pont, Wilmington DE) [18, 37].  

The success of color flow duplex sonography in imaging the calf veins has changed over the years. [9, 18, 19] Venography has confirms that calf involvement is the number one area of thrombus occurring 83% of the time [36]. Soleal vein thrombus was the most frequent site of deep vein thrombosis in which a fatal pulmonary thromboembolism was demonstrated on autopsy [9]. Sonography has shown to be more effective in distinguishing calf DVT than venography [26, 18, and 19].

The DVTrax 2.1 is a tool that shows the potential trajectories or tracks of a given patient's DVT (with associated probabilities), based on aggregated patient data.  For a given patient, the user inputs areas of observed thrombus to determine: (a) other areas to image today where, based on aggregated patient data, thrombus may be present; (b) other areas to image in the future. We think this tool can be used for quality assurance purposes for interpreting physicians and with additional multicenter data could aid in the decision on how to treat certain diagnosed thrombus. 

Conclusion:

Calf vein thrombus is found with above the knee thrombus (98.17%) of the time. Peroneal (63.61%), Posterior Tibial (50%), Gastrocnemius (36.91%) and Soleal (28.53%) veins should be routinely imaged as the rate of thrombus is significant.  The increased detection of calf deep vein thrombus and subsequent decision to treat may lower the incidence of thromboembolic events. It is very apparent in the research that those who routinely image calf veins and muscular calf veins recognize the amount of thrombus identified and that thrombus propagates.  Practice variation in the evaluation of the lower extremity for DVT does cause treatment uncertainty.  Standardization of the exam needs to be addressed from a multi-disciplined point of view.  The advancements in diagnostic equipment, along with developments in the technical application of ultrasound should be weighed in the discussion. We feel that there is evidence to support standardization across service lines for whole leg lower extremity venous testing to include imaging of the deep calf veins and muscular calf veins.