Version July 2025
INTRODUCTION —
Thrombosis of the lower extremity deep veins (figure 1) is termed deep venous thrombosis (DVT). Pulmonary embolism and, rarely, paradoxical embolism are complications of untreated DVT such that timely diagnosis of DVT is essential.
The approach described in this topic applies to nonpregnant adults. The evaluation of pregnant females with suspected DVT and the treatment of DVT are discussed separately. (See "Overview of the treatment of lower extremity deep vein thrombosis (DVT)" and "Deep vein thrombosis in pregnancy: Clinical presentation and diagnosis", section on 'Initial diagnostic evaluation'.)
EPIDEMIOLOGY —
DVT is common. In general, the incidence of venous thromboembolism (VTE), comprising DVT and pulmonary embolism, is approximately 1 per 1000 annually in adults [1,2].
Approximately two-thirds of VTE episodes manifest clinically as DVT.
The prevalence of DVT varies with the population studied (eg, inpatient, outpatient, surgical, medical) and underlying risk factor(s) (eg, hormonal contraception, surgery, cancer). These details are discussed separately. (See "Overview of the causes of venous thrombosis in adults".)
There is a slight male preponderance.
ADVERSE HEALTH OUTCOMES
Adverse health outcomes include:
●Pulmonary embolism, which can be fatal or lead to chronic pulmonary hypertension.
●Post-thrombotic syndrome, a syndrome of chronic venous insufficiency.
●Paradoxical embolization including stroke when patent foramen ovale is present.
(See "Clinical presentation and diagnostic evaluation of the nonpregnant adult with suspected acute pulmonary embolism" and "Epidemiology, pathogenesis, clinical manifestations and diagnosis of chronic thromboembolic pulmonary hypertension" and "Post-thrombotic (postphlebitic) syndrome in adults" and "Atrial septal abnormalities (PFO, ASD, and ASA) and risk of cerebral emboli in adults".)
CLINICAL PRESENTATION —
Features of lower extremity DVT are nonspecific, and some patients are asymptomatic (table 1).
History — DVT should be suspected in patients who present with leg swelling, pain, warmth, and erythema [3-5]. In one series, the sensitivity and specificity of these findings were [6]:
●Swelling or edema – 97 and 33 percent
●Pain – 86 and 19 percent
●Warmth – 72 and 48 percent
Symptoms are usually unilateral but can be bilateral.
In patients with isolated distal DVT, symptoms are confined to the calf while patients with proximal DVT may have calf or whole-leg symptoms. Iliac vein thrombosis may present with massive swelling of the proximal part of the leg and buttock pain. Phlegmasia is an uncommon form of massive proximal DVT that can be limb-threatening and is discussed separately. (See "Overview of the treatment of lower extremity deep vein thrombosis (DVT)", section on 'Special populations'.)
Risk factors may or may not be present (table 2) prompting the terms provoked (DVT with risk factors) and unprovoked (DVT without risk factors) (table 3). Risk factors for DVT are discussed separately. (See "Overview of the causes of venous thrombosis in adults".)
Physical examination — While the examination may be unrevealing, classic findings include the following:
●Unilateral edema or swelling with a difference in calf or thigh circumferences
●Unilateral warmth, tenderness, erythema
●Tenderness along the course of the involved major veins ("tender cords") (figure 1)
●Dilated superficial veins
●Local (eg, inguinal mass) or general signs of malignancy
Among these, a difference in calf circumference is the most useful finding [7].
Homans sign (calf pain on passive dorsiflexion of the foot) is unreliable for the presence of DVT.
Laboratory — Routine laboratory tests (eg, complete blood count, chemistries, liver function tests, coagulation studies) are not useful diagnostically but facilitate the differential diagnosis and if DVT is diagnosed, they may provide clues as to the underlying cause. (See "Evaluating adult patients with established venous thromboembolism for acquired and inherited risk factors", section on 'Laboratory tests and imaging'.)
Similarly, the D-dimer level is not diagnostic but can be used for pretest probability assessment to determine subsequent testing. (See 'Suspected first DVT' below and 'Measure D-dimer' below.)
Differential diagnosis — More than three-quarters of patients with suspected DVT have another cause for their symptoms (table 4) [8,9]. The following causes of the leg pain were identified in 160 consecutive patients with suspected DVT who had negative venograms [8]:
●Muscle strain, tear, or twisting injury to the leg – 40 percent
●Leg swelling in a paralyzed limb – 9 percent
●Lymphangitis or lymphatic obstruction – 7 percent
●Venous insufficiency – 7 percent
●Popliteal (Baker's) cyst – 5 percent
●Cellulitis – 3 percent
●Knee abnormality – 2 percent
●Unknown – 26 percent
There are no pathognomonic clinical features that reliably distinguish DVT from competing diagnoses since erythema, swelling, tenderness, and pain are common to many of these disorders. In addition, DVT may complicate such conditions. Thus, patients with these symptoms and signs often have compression ultrasonography (CUS) testing to exclude or diagnose concurrent DVT.
Findings that suggest specific disorders are the following:
●Calf muscle pull or tear – An inciting injury may be identified in the history, and there may be signs of bleeding on ultrasonography or bruising at the ankle. Rupture of the plantaris muscle tendon is characterized by acute calf pain and a popping sound. (See "Non-Achilles ankle tendinopathy".)
●Cellulitis – Bacterial cellulitis is a common condition of the leg, particularly in patients who have chronic leg swelling due to venous insufficiency or lymphedema. In bacterial cellulitis, the warmth and redness often skip areas and may be associated with constitutional symptoms including fever. Some patients with venous insufficiency develop a low-grade, nonbacterial cellulitis, which resembles infectious cellulitis, but without constitutional symptoms. Although fever increases the suspicion for cellulitis, it can also be present in DVT. (See "Cellulitis and skin abscess: Epidemiology, microbiology, clinical manifestations, and diagnosis".)
●Superficial thrombophlebitis – Superficial vein thrombophlebitis classically presents with palpable, tender superficial veins. However, since superficial thrombophlebitis is a risk factor for DVT, many of these individuals also undergo CUS to exclude DVT. (See "Superficial vein thrombosis and phlebitis of the lower extremity veins".)
●Venous valvular insufficiency – Venous insufficiency is a common cause of chronic unilateral leg edema (picture 1) and may be associated with varicose veins. (See "Diagnostic evaluation of lower extremity chronic venous disease" and "Clinical manifestations of lower extremity chronic venous disorders".)
●Lymphedema – Lymphedema is an important cause of chronic, rather than acute, edema of the extremities and is associated with nonpitting of the subcutaneous tissue due to fibrous accumulation over time. A history of pelvic surgery, malignancy, or radiation therapy is often present as the underlying etiology. (See "Clinical features, diagnosis, and staging of peripheral lymphedema".)
●Popliteal (Baker's) cyst – The majority of popliteal cysts are due to either distention of a bursa by fluid originating from a knee joint or posterior herniation of the joint capsule due to increased intra-articular pressure. A popliteal cyst that causes calf symptoms is usually leaking or has ruptured. It is often distinguished from DVT by posterior knee pain, knee stiffness, swelling or a mass behind the knee (especially with the knee in extension), and bruising around the ankle. However, calf symptoms are common; in addition, compression of the popliteal vein may cause leg swelling or secondary DVT. (See "Popliteal (Baker's) cyst".)
●Inflammatory pathology of the knee – Pain, inflammation, and swelling can accompany any knee joint pathology, which can be confused with a popliteal cyst or DVT. (See "Approach to the adult with unspecified knee pain".)
●Drug-induced edema or heart failure – Leg swelling is a side effect of some drugs, such as calcium channel blockers, and is commonly found in patients with congestive heart failure. The edema is usually bilateral but can be asymmetric if there is underlying venous pathology; signs of inflammation are not generally present. (See "Heart failure with preserved ejection fraction: Clinical manifestations and diagnosis", section on 'Clinical manifestations' and "Approach to diagnosis and evaluation of acute decompensated heart failure in adults", section on 'Clinical manifestations' and "Clinical manifestations and evaluation of edema in adults".)
SUSPECTED FIRST DVT —
Our approach incorporates the following (algorithm 1):
●Clinical assessment of the pretest probability (PTP). (See 'Initial approach (pretest probability)' below.)
●Selective D-dimer testing. (See 'Measure D-dimer' below.)
●Strategic use of compression ultrasonography (CUS) with Doppler for diagnosis (see 'Diagnostic compression ultrasonography (CUS)' below).
●Alternative imaging modalities (eg, computed tomographic [CT] venography [CTV] or magnetic resonance venography [MRV]) if CUS is unhelpful and suspicion for DVT remains. (See 'Alternative imaging' below.)
Our PTP-driven approach is in general consistent with several international guidelines and societies [10-13].
As an alternative, some clinicians proceed directly to CUS with Doppler with the caveat that CUS may be unnecessary for many patients. Diagnostic CUS is discussed separately. (See 'High-probability (likely)' below.)
Goal of diagnostic testing — The goal of diagnostic testing is to "rule in" (>85 percent posttest probability of DVT) or "rule out" DVT (<3 percent posttest probability of venous thromboembolism [VTE] in the next three months) with an acceptable level of certainty, thereby justifying instituting or withholding anticoagulant therapy, respectively [10].
Initial approach (pretest probability) — In patients with suspected first DVT, we estimate the PTP, typically using the Wells or modified Wells score (calculator 1) (table 5). Gestalt estimates are an alternative and may be supplemented by formal PTP assessment (see 'Wells and modified Wells score' below). Neither gestalt estimates nor prediction rules have been conclusively proven to be superior to the other.
While we typically prefer to assign patients to three risk categories (low; moderate; high [Wells]), assigning patients to two categories (unlikely; likely [modified Wells]) is also appropriate.
Subsequent diagnostic evaluation is determined by the assigned PTP as outlined in the sections below:
●(See 'Low or moderate (unlikely) probability' below.)
●(See 'High-probability (likely)' below.)
Wells and modified Wells score — Among the available PTP scoring systems, the best studied are the Wells and modified Wells scores [5,14-24].
●Wells score – Components of the Wells score are listed in the table (table 5) (calculator 1) [16]:
The total score in an individual patient denotes the following risk of DVT:
•≤0 points – Low probability (3 percent prevalence DVT) (see 'Low or moderate (unlikely) probability' below)
•1 to 2 points – Moderate probability (17 percent prevalence DVT) (see 'Low or moderate (unlikely) probability' below)
•3 to 8 points – High probability (50 to 75 percent prevalence DVT) (see 'High-probability (likely)' below)
The scoring system may perform differently in select populations [7,25,26]. For example, one prospective study of 1295 patients with suspected DVT in a primary care setting reported a higher rate of DVT in patients with a low probability of DVT (12 percent, as opposed to the expected 3 percent) [25]. A meta-analysis of 51 studies also reported reduced performance of the score in patients who were older, had a prior DVT, or had comorbidities other than those included on the score [7].
●Modified Wells score – The modified Wells score contains all the components of the original Wells score with one additional point given to those with a history of previously documented DVT (table 5) [27]. The modified score classifies patients according to whether DVT is likely (a score ≥2) or unlikely (a score ≤1). Use of the modified Wells score was best studied in an algorithm that included measurement of D-dimer [27], which is discussed below. (See 'Measure D-dimer' below.)
Low or moderate (unlikely) probability
Measure D-dimer — In patients with a low or moderate PTP for first lower extremity DVT (ie, approximately 3 and 17 percent probability, respectively; ≤2 on Wells score or <2 on modified Wells score) (calculator 1) (table 5), we obtain a high-sensitivity D-dimer level. Use of PTP and D-dimer is best validated in outpatients with leg symptoms. (See 'Initial approach (pretest probability)' above.)
D-dimer should not be done if it is expected to be positive due to another condition (table 6); in such cases, CUS should be performed.
As an alternative, especially in those with a moderate PTP for DVT, some experts proceed directly to CUS without D-dimer testing. The decision is influenced by factors including CUS availability and whether CUS testing will evaluate other diagnoses (eg, Baker's cyst).
Interpretation and subsequent steps — The subsequent approach is the following:
●Normal D-dimer level (<500 ng/mL) – For patients with a low or moderate PTP in whom a high-sensitivity D-dimer level is normal, we consider DVT excluded and no further testing is needed. (See 'Measure D-dimer' above.)
If a moderate-sensitivity assay was performed, a normal D-dimer level is sufficient for the exclusion of DVT only in those with a low suspicion for DVT but not for those with a moderate suspicion, in which case CUS with Doppler should be performed.
●Positive D-dimer level (≥500 ng/mL) – Patients with a low or moderate PTP in whom a high-sensitivity D-dimer is positive, we perform whole-leg or proximal CUS with Doppler. Choosing whole-leg or proximal CUS is discussed below (see 'Diagnostic compression ultrasonography (CUS)' below and 'Choosing proximal or whole-leg ultrasonography' below):
•Positive CUS – For patients with a positive CUS, DVT is diagnosed.
•Negative CUS – For patients with a negative whole-leg or proximal CUS, the approach differs depending on the PTP:
-Low suspicion for DVT – When the suspicion for DVT is low and CUS is negative, we consider DVT as excluded and no further testing is required.
-Moderate suspicion for DVT – The approach differs depending on whether whole-leg or proximal CUS is performed.
If whole-leg ultrasonography is negative (ie, neither proximal nor distal DVT is identified), we consider DVT excluded and do not perform additional testing.
If proximal CUS is negative, distal DVT may have potentially been missed. Options include whole-leg CUS with Doppler to evaluate the distal veins or repeat proximal CUS with Doppler at seven days for two weeks (typically off anticoagulation) to detect possible extension of distal DVT into the proximal veins. If whole-leg or repeat proximal CUS is negative, we consider DVT excluded and no further testing is needed. (See 'Choosing proximal or whole-leg ultrasonography' below.)
•Nondiagnostic CUS – Further investigation based upon nondiagnostic findings are discussed below. (See 'Ultrasonography interpretation' below.)
Performance characteristics (D-dimer) — Most institutions use a "high-sensitivity" assay. However, cutoff values and assay type vary among laboratories. Thus, clinicians should be aware of which assay is being used in their institution. We prefer "high-sensitivity" assays since they can reliably exclude DVT in a patient with low or moderate PTP while "moderate-sensitivity" assays are only useful for excluding DVT in those with low PTP.
●High-sensitivity assays – One meta-analysis of 217 studies reported the highest sensitivity associated with the following D-dimer assays [10]:
•Enzyme-linked immunofluorescent assays (ELISA; 96 percent)
•Microplate ELISA (94 percent)
•Semiquantitative or immunoturbidimetric assays (93 percent)
●Moderate-sensitivity assays – Whole blood D-dimer and latex semiquantitative assays are less sensitive (83 to 85 percent) [10]. Although moderate-sensitivity D-dimer assays are more specific than high-sensitivity assays, they have a lower negative predictive value.
Studies consistently report a high sensitivity and poor specificity for D-dimer in patients with acute DVT [10,28-34]. D-dimer level is elevated in nearly all patients with acute DVT (ie, it is highly sensitive). However, it is nonspecific since elevated levels are found in many other conditions (table 6) [10,35-46]. Using a validated high-sensitivity D-dimer assay, a negative result (eg, <500 ng/mL) is useful for excluding DVT in those with a low or moderate PTP for DVT. However, a negative test is obtained in only 30 percent of outpatients (lower in inpatients or if there has been a previous VTE). A positive result (eg, ≥500 ng/mL) is not diagnostic and indicates the need for CUS with Doppler.
Data that support using D-dimer level in conjunction with PTP or CUS for excluding DVT in patients with low or moderate suspicion for DVT include the following [27,29,33,34,36,39,43,47-63]:
●In one trial, 1096 patients with suspected DVT were classified according to the modified Wells criteria as DVT "likely" or "unlikely" [27] (table 5) (see 'Wells and modified Wells score' above). The patients were then randomly assigned to undergo routine CUS or to CUS only if they had an unlikely PTP and a positive D-dimer or if they had a likely PTP (D-dimer group). D-dimer testing resulted in a significant reduction in the use of CUS (0.78 versus 1.34 tests per patient). For patients in whom DVT was excluded by the initial diagnostic strategy, the frequency of VTE during three months of follow-up was similar (0.4 versus 1.4 percent in the D-dimer and CUS groups, respectively).
●Several studies have reported that a D-dimer level <500 ng/mL also removes the need for serial proximal CUS in those with an initial negative CUS [64-66]. As an example, a study of 810 patients with suspected DVT (low intermediate and high PTP) and negative CUS on initial testing reported a similar six-month rate of thromboembolism in patients with a negative D-dimer and those with a negative repeat ultrasonography at one week (2.1 versus 1.3 percent) [66].
Adjusted D-dimer approaches have also been described:
●Age-adjusted – D-dimer values rise with age, further hampering specificity in older patients [36,67-70]. Using a higher D-dimer cutoff in older patients improves its diagnostic utility and specificity [69-72]. While age-adjusted D-dimer cutoffs are validated, use is variable among institutions.
One meta-analysis of 13 studies (12,497 patients) compared the specificity of conventional cutoff D-dimer values (<500 ng/mL) with age-adjusted values (defined as age [years] x 10 ng/mL for patients >50 years) [70]. Higher specificities were reported for age-adjusted cutoff values (age 51 to 60 years: 63 versus 58 percent; 61 to 70 years: 50 versus 39 percent; 71 to 80 years: 44 versus 24 percent; >80 years: 35 versus 15 percent).
●"PTP-adjusted" – A "PTP-adjusted" approach to D-dimer interpretation has been prospectively studied in patients with suspected DVT [73]. DVT was excluded when D-dimer levels were <1000 ng/mL (patients with low PTP) or <500 ng/mL (patients with moderate PTP). This strategy reduced the need for CUS by 8 percent in outpatients suspected of having DVT and by 21 percent in those with low PTP.
Another study of 1508 patients used a similar strategy to exclude DVT. Only 0.6 percent of those in whom DVT was excluded developed DVT during follow-up. [54]. It was estimated that this strategy reduced the need for imaging by 47 percent when compared with a traditional strategy (1.36 to 0.72 scans per patient). While encouraging, further study is needed to validate this approach.
High-probability (likely)
Diagnostic compression ultrasonography (CUS) — In patients with a high PTP for first lower extremity DVT (ie, 50 to 75 percent prevalence of DVT) (calculator 1) (table 5), we obtain proximal or whole-leg CUS with Doppler. Typically, both legs are imaged. Unlike patients with a low or moderate PTP, D-dimer cannot be reliably used to exclude DVT since the negative predictive value is lower in this population. Factors that influence whether proximal or whole-leg CUS is chosen are discussed separately. (See 'Choosing proximal or whole-leg ultrasonography' below.)
The subsequent approach is the following:
●Positive ultrasonography – Patients with a positive CUS are diagnosed with DVT.
●Negative CUS – For patients with a high suspicion for DVT in whom CUS is negative, subsequent testing should be individualized and partly driven by the suspected location of DVT. (See 'Ultrasonography interpretation' below.)
•Iliac vein thrombosis suspected – If iliac vein thrombosis is suspected (eg, massive upper leg swelling, buttock pain), we obtain Doppler ultrasonography of the iliac vein. If not feasible or inconclusive, alternative imaging modalities may be useful (eg, contrast-enhanced CTV or non-contrast MRV). (See 'Alternative imaging' below.)
•Distal DVT suspected – If proximal CUS is negative, then distal DVT may have been missed. If distal DVT is suspected, options include repeat proximal CUS at three and seven days (typically off anticoagulation) or whole-leg CUS. If not suspected to be elevated due to another condition (table 6), high sensitivity D-dimer level measurement may be useful, if negative. A positive D-dimer is not useful.
•Nondiagnostic CUS – Options for a nondiagnostic CUS are discussed below, but typically, alternate imaging is required. (See 'Alternative imaging' below and 'Ultrasonography interpretation' below.)
In general, if one or more of these tests are negative, then DVT is considered excluded and no further testing is needed. Patients should be followed clinically and investigations performed to find other etiologies. (See "Overview of the causes of venous thrombosis in adults", section on 'Anatomic risk factors for deep vein thrombosis'.)
Choosing proximal or whole-leg ultrasonography — Proximal vein CUS detects thrombus in the proximal veins only (figure 1). Whole-leg ultrasonography additionally examines the calf veins to identify isolated distal DVT; whole-leg CUS is our preference and is increasingly becoming the preferred approach in the United States [10-12,74].
Factors affecting the type of CUS used include the following:
●Institutional practice – In practice, the choice between proximal or whole-leg ultrasonography is often institution-specific. It is important that clinicians know which test is performed at their institution and are familiar with its advantages and disadvantages. (See 'Initial approach (pretest probability)' above and 'Measure D-dimer' above.)
●Clinician preference – Preference varies:
•Many experts prefer whole-leg CUS since it can detect both distal and proximal DVT. Unlike proximal vein CUS, a negative whole-leg CUS excludes all cases of DVT, and may, therefore, reduce the need, time, and expense of additional diagnostic testing or repeat proximal CUS [75]. However, whole-leg CUS may identify DVT that does not necessarily need to be treated (some cases of distal DVT resolve spontaneously). Consequently, some clinicians limit the performance of whole-leg ultrasonography to select subgroups (eg, those with prominent calf symptoms and negative proximal CUS).
•Some clinicians prefer proximal vein CUS since it identifies clinically actionable DVT. However, a negative study may miss distal DVT. (See "Overview of the treatment of lower extremity deep vein thrombosis (DVT)", section on 'Distal DVT' and "Overview of the treatment of lower extremity deep vein thrombosis (DVT)", section on 'Low risk of extension: Serial ultrasound'.)
●Time and technical issues – Proximal CUS can usually be performed within 5 to 10 minutes and is technically easier to perform, and interpretation is less operator-dependent. In contrast, whole-leg CUS usually takes 15 to 20 minutes and is more difficult to perform and interpretation is operator-dependent.
Ultrasonography interpretation — Interpretation of CUS in patients with a first suspected DVT is described in this section. Interpreting CUS in patients with suspected recurrent DVT is discussed separately. (See 'Ultrasonography interpretation' below.)
●Positive – Using ultrasound probe pressure, the presence of thrombus is diagnosed by demonstrating noncompressibility of the imaged vein. Veins that can be assessed for compressibility are the proximal veins (eg, the common femoral, femoral, and popliteal veins). The distal veins are less compressible (eg, peroneal, posterior and anterior tibial, and muscular veins), followed by the iliac veins (figure 1).
The term "free-floating" DVT is sometimes applied to thrombus that appears poorly adherent to the adjacent vessel wall. There are no strict diagnostic criteria for this phenomenon. It is reasonable to suggest that the risk of progression and embolization is similar to adherent thrombus and that it should be treated as such.
●Negative – A negative study is one that demonstrates full compressibility of all imaged veins.
●Nondiagnostic – A nondiagnostic study is one where there is uncertainty about whether DVT is present or absent.
Nondiagnostic findings are less common in outpatients compared with inpatients (<5 percent of outpatients expected to have nondiagnostic findings). Nondiagnostic findings are also less common when imaging the proximal compared with the distal veins.
There are three main reasons for a nondiagnostic examination, which often overlap:
•First, there may be difficulty visualizing the deep veins because of severe obesity, edema, recent surgery or trauma, skin lesions, contractures, or leg casts (ie, technically limited studies) [76].
•Second, although the deep veins are well visualized, small (eg, <5 cm long or <2 mm wide) or atypical appearing abnormalities of uncertain significance may be identified [77].
•Third, when thrombus is present in patients with previous DVT, it is often difficult to assess if it is new (acute) or old (residual thrombosis can persist indefinitely), the details of which are discussed separately. (See 'Ultrasonography interpretation' below.)
In those with nondiagnostic studies, we individualize further investigation. Choosing another imaging study depends upon why the CUS is considered nondiagnostic, the extent and position of the venous segment that is nondiagnostic (eg, distal or proximal veins), clinical PTP, results of D-dimer testing, and the clinician's overall assessment of the risk associated with undiagnosed DVT. Options include repeat proximal CUS at seven days (eg, distal veins were poorly imaged in a patient with high clinical suspicion) or alternate imaging (eg, technical limitations that are not quickly resolvable, such as obesity, contractures, or casting; distinguishing old from new thrombus). (See 'Alternative imaging' below.)
Performance characteristics (CUS) — Lower extremity CUS is the diagnostic test of choice for the diagnosis of proximal vein DVT due to its high sensitivity and specificity (>95 percent each) [78-80]. However, CUS is less sensitive for the detection of calf vein and iliac vein thrombus since these veins are less readily compressed (particularly calf veins) [78-81].
Supportive data for proximal and whole-leg CUS for the diagnosis of DVT include the following:
●Proximal CUS – Several randomized trials and meta-analyses have consistently reported that proximal CUS is a sensitive test for proximal DVT compared with contrast venography (95 to 100 percent) [10,82-87].
CUS also has a high negative predictive value for the exclusion of DVT. As examples:
•For patients with a low probability for DVT, a single proximal CUS excludes the diagnosis, with rates of VTE during three months follow-up of about 0.5 percent [10].
•For patients with a moderate or high probability of DVT who have a negative proximal CUS at initial presentation, approximately 2 percent have a proximal DVT when retested seven days later [88].
Serial proximal DVT (at seven days) has been shown in randomized trials to perform similarly to whole-leg CUS for the detection of clinically significant DVT when anticoagulants are withheld. In these trials, the rate of DVT in the subsequent three months has been reported to be approximately 0.5 to 2 percent [85,89,90].
●Whole-leg CUS – Data that support the high sensitivity of whole-leg CUS in patients with suspected DVT include several randomized studies and meta-analyses (>99 percent) [84,90-97]. One meta-analysis of seven studies also reported that the three-month rate of VTE in patients who had anticoagulants withheld after a single negative whole-leg ultrasound is very low (0.6 percent; 95% CI 0.25-0.89) [97]. Rates were 0.3 percent for patients with low-probability DVT, 0.8 percent for those with moderate-probability DVT, and 2.5 percent for those with high-probability DVT.
Sometimes CUS is combined with venous Doppler (commonly in the United States). The addition of color flow Doppler does not improve CUS sensitivity, but it can provide supportive evidence of thrombosis and help identify calf veins [98,99]. Doppler ultrasonography alone should not be used since it is less sensitive and less specific than CUS [74].
This topic does not discuss the accuracy of point-of-care ultrasonography, which is not recommended for diagnosis unless the situation is urgent and should be performed by clinicians trained in its use. (See "Indications for bedside ultrasonography in the critically ill adult patient", section on 'Vascular ultrasonography'.)
PATIENTS WITH A HISTORY OF DVT —
A previous episode of DVT is a known risk factor for a recurrent event. Patients with suspected recurrence present with similar symptoms and signs as those with a first event. (See 'Clinical presentation' above.)
Suspected contralateral recurrence — For the diagnosis of recurrent DVT on the contralateral leg, we follow the same approach as for patients with a suspected first event. (See 'Suspected first DVT' above.)
Suspected ipsilateral recurrence
Initial evaluation — For most patients with suspected recurrent ipsilateral DVT, we perform compression ultrasonography (CUS; proximal or whole-leg) based upon a presumed high probability of DVT [10]. Interpretation of CUS is different from that described in patients with a first event. (See 'Ultrasonography interpretation' below.)
As an alternative, using an approach similar to that described for first suspected DVT is also appropriate. Prediction scoring systems such as the Wells criteria (table 5) (calculator 1) or the modified Wells [60,100] have been used to reduce the number of unnecessary ultrasound scans in those with suspected recurrence as well as guide management when scans are nondiagnostic. Evidence suggests that, similar to patients with a first suspected event, a negative D-dimer level (eg, <500 ng/mL) is a sensitive diagnostic test and is therefore useful in excluding recurrence; however, when compared with patients with a first suspected event, the specificity is lower and fewer patients present with a negative D-dimer [27,100,101]. As an example, a prospective study of 105 patients with suspected recurrent DVT reported that D-dimer had a sensitivity of 97 percent, specificity of 30 percent, and negative predictive value of 95 percent [100]. However, the D-dimer was negative in only 17 percent of patients.
Ultrasonography interpretation — Thrombus resolves slowly with time, with a decrease in residual venous diameter (RVD) of about 50 to 60 percent over the first three months in the common femoral and popliteal veins [102]. About 80 percent of proximal ultrasound studies remain abnormal at three months, and about 25 to 50 percent are abnormal at one year [102-104]. Thus, when an abnormality is identified in patients with suspected recurrence, it may be difficult to determine whether it is due to old or new thrombus. Consequently, the specificity of ultrasonography for recurrence is lower than in patients with suspected first DVT. Although the specificity is improved when a previous ultrasound documents the extent of residual thrombosis, interobserver agreement remains variable and limited [105].
Previous ultrasonography — When there is a previous CUS for comparison, we and others agree that the following generally applies [10]:
●A positive CUS for acute thrombosis is one that demonstrates any one of the following:
•Evidence of new noncompressibility of the proximal vein – This feature should be considered diagnostic of acute DVT, although only a minority of recurrent DVT are identified this way (10 to 20 percent) [102,106]. (See "Overview of the treatment of lower extremity deep vein thrombosis (DVT)".)
•Evidence of an increase in the RVD ≥4 mm and/or substantial (eg, >10 cm) thrombus extension – These features are considered by most physicians as recurrent DVT. Although the optimal cut-off value for increase in RVD is uncertain, studies have reported that the specificity of an increase in RVD ≥2 mm was 95 percent and ≥4 mm was 100 percent [102,107,108].
●A negative CUS for acute thrombosis is one in which all venous segments are fully compressible or one with <2 mm increase in RVD of both the popliteal and common femoral vein compared with previous CUS.
•If whole-leg ultrasonography is negative for proximal and distal DVT, no further testing is required. If proximal CUS is negative, patients should undergo repeat proximal CUS at seven days or whole-leg ultrasonography or, rarely, high-sensitivity D-dimer level measurement (if not already performed). If any of these are negative, then no further testing is necessary.
•Patients with an increase in RVD <2 mm are unlikely to have recurrent DVT and should have repeat testing with proximal CUS at seven days. A stable or improved repeat CUS is not considered as new thrombus whereas evidence of growing thrombus is an indication for treatment.
The safety of diagnostic strategies that use RVD measurements on CUS to withhold anticoagulation when the CUS is compressible or reveals an increase in RVD <2 mm includes the following:
-A prospective study of 205 patients with suspected recurrent ipsilateral DVT compared proximal CUS findings with a CUS that had been performed after 3 to 12 months of anticoagulation [107]. Patients with a negative CUS (ie, noncompressibility or increased RVD <2 mm; n = 153) had serial testing at two and seven days, during which time anticoagulation was withheld; CUS converted to positive in 3 of the 153 patients. In the remaining 149 patients who remained off anticoagulant therapy, two (1.3 percent) had venous thromboembolism (VTE) during the next six months.
-In a similar study of 284 patients in whom recurrent DVT had been excluded using the same criteria and in whom anticoagulation was withheld, 3 percent developed VTE over the next three months [108].
-In a prospective study of 75 patients with suspected recurrence, when there was a negative D-dimer (<500 ng/mL) in addition to these ultrasound criteria, no patients developed VTE over the next three months [109].
●A nondiagnostic study is one with an increase in RVD ≥2 mm and <4 mm or an extension in thrombus length that is <10 cm. Management of nondiagnostic studies should be individualized and involve an assessment similar to that described above for high-probability first suspected DVT (eg, pretest probability [PTP] assessment, D-dimer, serial CUS, or CT or magnetic resonance venography). (See 'Interpretation and subsequent steps' above.)
No previous ultrasonography — In the absence of previous ultrasonography for comparison, the following generally applies:
●A CUS that demonstrates noncompressibility of the popliteal or common femoral vein indicates that there is either new or old thrombosis. New thrombosis is more likely if they have one or more of the following:
•Extensive thrombus
•Limited compressibility of the imaged vein
•Positive D-dimer, particularly if markedly elevated (eg, >2000 ng/mL)
•High PTP
●If the overall assessment is that acute thrombosis is likely, patients are generally treated. If the overall assessment is that acute thrombosis is unlikely, anticoagulation can be withheld while serial ultrasonography is performed (eg, seven days).
●A negative ultrasonography scan is one where all veins are fully compressible.
●Management of nondiagnostic studies should be individualized and involve an assessment similar to that described above. (See 'Interpretation and subsequent steps' above.)
ALTERNATIVE IMAGING —
For patients with suspected first or recurrent DVT contrast-enhanced CT venography (CTV) and magnetic resonance venography (MRV) are rarely used unless there is uncertainty after compression ultrasonography (CUS) has been performed. These diagnostic tests are considered less well-validated, have associated complications of radiation (eg, CTV) and contrast exposure, and are more costly. However, they may identify obstructing lesions or other etiologies for the patient's symptoms.
While widely used in the past, ascending contrast venography, which was the gold standard for DVT diagnosis, and impedance plethysmography are now essentially obsolete.
The technical aspects and diagnostic value of these modalities are discussed below:
●CTV – Thrombus is usually identified on CTV by demonstrating a filling defect with contrast-enhancement. The major disadvantage of CTV is the potential for contrast-related complications or contraindications. (See "Contrast-associated and contrast-induced acute kidney injury: Clinical features, diagnosis, and management" and "Prevention of contrast-associated acute kidney injury related to angiography" and "Patient evaluation prior to oral or iodinated intravenous contrast for computed tomography", section on 'Introduction'.)
CTV can be performed either as direct CTV using unilateral or bilateral venous injection of iodinated contrast into a pedal vein or, more commonly, as indirect CTV using an antecubital vein for contrast injection and delayed imaging acquisition to obtain optimal deep vein opacification [110,111].
There is little evidence to support the use of CTV to diagnose DVT except when CUS is nondiagnostic. Sensitivity and specificity data are derived from studies that combined indirect CTV with CT pulmonary angiography (CTPA) in patients with suspected pulmonary embolism [39,112-115]. Although sensitivities as high as 95 percent for the diagnosis of femoropopliteal thrombus have been reported, data are derived from small case series. (See "Clinical presentation and diagnostic evaluation of the nonpregnant adult with suspected acute pulmonary embolism", section on 'CTPA'.)
●Magnetic resonance imaging – Gadolinium-enhanced MRV and magnetic resonance direct thrombus imaging (MRDTI) can be used to diagnose DVT. Occasionally, thrombus, which appears as a filling defect on MRV, can be identified without contrast, but this is not the preferred method.
•MRV is less well-studied than CTV. Small case series report sensitivities >95 percent for the diagnosis of lower extremity proximal DVT [116-120].
•MRDTI is an evolving technique that can diagnose lower extremity DVT based upon the altered characteristics of red blood cell methemoglobin within the imaged vein. MRDTI does not involve gadolinium contrast as the methemoglobin gives a spontaneous T1 signal. Another advantage is that it takes only 5 to 10 minutes to make a whole-leg image. However, MRDTI is not universally available.
Small prospective studies suggest MRDTI can distinguish new from old thrombus and may therefore be a valuable tool for the diagnosis of ipsilateral recurrence [121-123]. In a prospective study of recurrent ipsilateral DVT, <2 percent of patients who were MRDTI-negative developed DVT (off anticoagulation) during follow-up [122].
●Ascending contrast venography – Ascending contrast venography is performed by injecting iodinated contrast into a dorsal foot vein to outline the deep venous system of the lower extremity. An intraluminal defect that is present in more than one view is diagnostic of DVT. It is invasive, expensive, technically difficult to perform (the dorsal vein cannot be cannulated in 5 percent of cases), difficult to interpret, and associated with contrast-related complications [8,86,124-128]. (See "Allergy evaluation of immediate hypersensitivity reactions to radiocontrast media" and "Radiocontrast hypersensitivity: Nonimmediate (delayed) reactions" and "Diagnosis and treatment of an acute reaction to a radiologic contrast agent".)
●Intravascular ultrasonography – This modality is rarely used and reserved mostly for guidance during interventional procedures for venous obstruction (eg, May-Thurner syndrome). (See "Overview of iliocaval venous obstruction", section on 'Venography and intravascular ultrasound' and "May-Thurner syndrome", section on 'Intravascular ultrasound'.)
DIAGNOSIS —
Noncompressibility of the imaged vein on compression ultrasonography (CUS) diagnoses a first-time DVT with a high degree of certainty (ie, >85 percent posttest probability).
A negative whole-leg CUS also reliably excludes DVT; <3 percent of patients develop venous thromboembolism three months after a negative CUS despite a high pretest probability.
A filling defect on contrast-enhanced CT venography or non-contrast magnetic resonance venography or abnormal flow on iliac vein Doppler ultrasonography are also diagnostic of DVT.
The diagnosis of recurrent DVT is complex and discussed above. (See 'Patients with a history of DVT' above.)
EMPIRIC ANTICOAGULATION —
The decision to empirically provide anticoagulant therapy while diagnostic testing is being performed depends upon the clinical suspicion for DVT and how soon diagnostic testing can be completed. Empiric anticoagulation is discussed separately. (See "Acute pulmonary embolism in adults: Treatment overview and prognosis", section on 'Empiric anticoagulation or thrombolysis'.)
SPECIAL POPULATIONS —
The presentations and diagnosis of DVT in special populations are discussed separately:
●Pregnancy (see "Deep vein thrombosis in pregnancy: Clinical presentation and diagnosis")
●Phlegmasia (see "Phlegmasia and venous gangrene")
●Upper extremity DVT (see "Primary (spontaneous) upper extremity deep vein thrombosis")
●May-Thurner syndrome (see "May-Thurner syndrome")
SOCIETY GUIDELINE LINKS —
Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Superficial vein thrombosis, deep vein thrombosis, and pulmonary embolism".)
INFORMATION FOR PATIENTS —
UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)
●Basics topics (see "Patient education: Deep vein thrombosis (blood clot in the leg) (The Basics)" and "Patient education: Duplex ultrasound (The Basics)")
●Beyond the Basics topics (see "Patient education: Deep vein thrombosis (DVT) (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Epidemiology – Thrombosis of the lower extremity deep veins (figure 1) is termed deep venous thrombosis (DVT). The incidence of venous thromboembolism (VTE), comprising DVT and pulmonary embolism, is approximately 1 per 1000 annually in adults. Approximately two-thirds of VTE episodes manifest as DVT. However, the prevalence varies with the population studied (eg, inpatient, outpatient, surgical, medical) and underlying risk factor(s) (eg, hormonal contraception, surgery, cancer) (table 2). (See 'Introduction' above and 'Epidemiology' above.)
●Clinical manifestations – Manifestations include the following (see 'Clinical presentation' above):
•Common features are unilateral leg swelling, pain, tenderness, warmth, and erythema. Bilateral DVT is uncommon (table 1). Manifestations are localized to the calf with distal DVT and involve the whole-leg with proximal DVT. Massive swelling of the proximal leg and buttock suggests iliac vein DVT/phlegmasia; reduced or absent pulses suggest associated arterial insufficiency. Asymmetry in the circumference of the calf or thigh is a useful finding. Tenderness along the course of the involved major veins ("tender cords") may also be evident (figure 1). Homans sign (calf pain on passive dorsiflexion of the foot) is unreliable. Occasionally, patients are asymptomatic or have limited findings on examination. (See 'History' above and 'Physical examination' above.)
•Complete blood count, chemistries, liver function tests, and coagulation studies should be obtained, but findings are variable and nonspecific for DVT. A D-dimer level should be measured to facilitate pretest probability (PTP) assessment. However, an elevated D-dimer is nonspecific and not diagnostic of DVT (table 6). (See 'Laboratory' above.)
●Differential diagnosis – The differential diagnosis is listed in the table (table 4). Clinical features may support one or more competing diagnoses. However, symptoms overlap and DVT can complicate these disorders. History, examination, and compression ultrasonography (CUS), if indicated, usually distinguish these conditions from DVT. (See 'Differential diagnosis' above.)
●Initial probability assessment – When DVT is suspected (first event), PTP should be estimated as low, moderate, or high (algorithm 1). This determines the next level of testing. However, some experts obtain ultrasonography in all patients without probability assessment.
Use of PTP and D-dimer is best validated in outpatients with leg symptoms. Gestalt assessment is a reasonable alternative. (See 'Initial approach (pretest probability)' above.)
●Low or moderate probability – In patients with a low or moderate PTP for DVT, we perform a high-sensitivity D-dimer level; a normal high-sensitivity D-dimer level (eg, <500ng/mL) excludes DVT. (See 'Low or moderate (unlikely) probability' above and 'Measure D-dimer' above.)
If the D-dimer is elevated (eg, ≥500 ng/mL), we perform CUS. We prefer whole-leg rather than proximal vein CUS; however, institutional practice varies.
•If CUS is positive, DVT is diagnosed.
•If CUS is negative and probability for DVT is low, DVT is excluded.
•If CUS is negative and probability for DVT is moderate, we perform whole-leg CUS if not done or repeat proximal vein CUS weekly for two weeks.
●High probability – Empiric anticoagulation is appropriate in patients with high PTP for DVT when testing is in progress if bleeding risk is not high. For patients with a high PTP for DVT, we perform CUS. We prefer whole-leg CUS rather than proximal vein CUS.
If DVT is not identified on CUS, further testing is required. Options include whole-leg CUS if not already performed, repeat CUS weekly for two weeks, iliac vein Doppler, and alternative imaging, such as CT venography (CTV) or magnetic resonance venography (MRV).
Recurrence – Patients with a prior history of DVT may be considered to have a moderate to high PTP for DVT.
•For most patients with suspected ipsilateral DVT recurrence, we perform CUS. Comparison with prior CUS images may be required for diagnosis. In the absence of a previous CUS, magnetic resonance direct thrombus imaging may be useful to distinguish old from new thrombus. As an alternative, using an approach like that described for first suspected DVT is also reasonable. (See 'Suspected ipsilateral recurrence' above.)
•Patients with symptoms in the contralateral leg may be evaluated using the same approach as patients with a suspected first event. (See 'Patients with a history of DVT' above and 'Suspected contralateral recurrence' above.)
●Alternative imaging – Contrast-enhanced CTV and MRV are rarely used unless there is diagnostic uncertainty after CUS has been performed. These diagnostic tests are considered less well-validated, have associated complications of radiation (eg, CTV) and contrast exposure, and are more costly. However, they may identify obstructing lesions or other etiologies for the patient's symptoms. (See 'Alternative imaging' above.)
●Diagnosis – Noncompressibility of the imaged vein on CUS diagnoses a first-time DVT with a high degree of certainty (ie, >85 percent posttest probability).
A negative whole-leg CUS also reliably excludes DVT; <3 percent of patients develop VTE three months after a negative CUS despite a high PTP.
A filling defect on contrast-enhanced CTV or non-contrast MRV or abnormal flow on iliac vein Doppler ultrasonography are also diagnostic of DVT.
ACKNOWLEDGMENTS —
The UpToDate editorial staff acknowledges Brydon Grant, MD, who contributed to earlier versions of this topic review.
The UpToDate editorial staff also acknowledges Clive Kearon, MB, MRCP(I), FRCP(C), PhD, now deceased, who contributed to earlier versions of this topic review.
