Compression ultrasound and duplex Doppler diagnosis of deep vein thrombosis — from Wells score to anticoagulation decision
Deep vein thrombosis affects an estimated 1–2 per 1,000 adults annually and is a leading preventable cause of in-hospital mortality via pulmonary embolism (PE). Because DVT symptoms — leg swelling, pain, warmth, erythema — are nonspecific and mimicked by cellulitis, muscle strain, and Baker's cyst rupture, clinicians use a validated clinical prediction rule, the Wells score, to convert a jumble of history and exam findings into a reproducible pretest probability category that calibrates how imaging and D-dimer results should be interpreted.
The original Wells score assigns one point each to: active cancer (treatment within 6 months or palliative), paralysis/paresis or recent plaster immobilization of the lower extremity, recently bedridden ≥3 days or major surgery within 12 weeks, localized tenderness along the deep venous system, entire leg swollen, calf swelling >3 cm compared to the asymptomatic leg (measured 10 cm below tibial tuberosity), pitting edema confined to the symptomatic leg, collateral superficial veins (non-varicose), and previously documented DVT. Two points are subtracted if an alternative diagnosis is at least as likely as DVT.
The resulting integer score is mapped to pretest probability. The original three-tier model: score ≤0 = low probability (~5% prevalence), 1–2 = moderate (~17%), ≥3 = high (~53%). A simplified two-tier model, validated for outpatient use with D-dimer, collapses this to "DVT unlikely" (≤1) versus "DVT likely" (≥2), streamlining the decision to order D-dimer versus proceed directly to ultrasound.
Pretest probability is not a formality — it is the Bayesian prior that determines what a negative or positive test actually means. In a low-probability patient, a negative high-sensitivity D-dimer (which has a negative predictive value approaching 99% in this setting) is sufficient to exclude DVT without any imaging at all, sparing the patient an ultrasound.
In a moderate or high-probability patient, D-dimer alone is insufficient because its specificity is poor (elevated in pregnancy, malignancy, recent surgery, infection, and simply older age) — these patients proceed directly to compression ultrasound regardless of D-dimer result. If ultrasound is negative but pretest probability was high, many protocols recommend a repeat scan in 5–7 days to catch a thrombus that was too small or too distal to detect on the first pass, since isolated calf-vein clot can be missed on a single study before it propagates proximally.
Wells score interpretation and paired action: Score ≤0 (Low, two-tier: ≤1) — DVT unlikely; D-dimer alone can exclude DVT if negative. Score 1–2 (Moderate) — D-dimer plus ultrasound; a negative D-dimer alone is not sufficiently reassuring. Score ≥3 (High, two-tier: ≥2) — Proceed directly to compression ultrasound; do not rely on D-dimer to exclude disease.
Compression ultrasonography (CUS) is the diagnostic workhorse for DVT: a linear high-frequency transducer (typically 5–10 MHz) is used to visualize the vein in transverse cross-section, and gentle, firm, direct pressure is applied through the probe every 2 cm along the venous course, from the common femoral vein down through the popliteal vein and, where feasible, the calf veins. The single most important sign in the entire examination is startlingly simple: does the vein wall coapt fully when compressed?
The vein is scanned in the transverse (short-axis) plane, which is far more sensitive to loss of compressibility than the longitudinal plane because the sonographer can watch the entire circular lumen collapse symmetrically. The probe is angled directly perpendicular to the skin and pressed straight down — enough to visibly deform the adjacent artery slightly, confirming adequate pressure was applied — while the artery, which is thick-walled and pulsatile, should never fully collapse.
Scanning proceeds systematically: common femoral vein (at the saphenofemoral junction), femoral vein along the adductor canal, popliteal vein in the popliteal fossa (patient often prone or with knee flexed), and, in a complete study, the paired posterior tibial and peroneal veins of the calf. Each segment is compressed at 1–2 cm intervals; failure to fully coapt the anterior and posterior vein walls at any point is diagnostic of thrombus at that level.
A normal, patent vein is a thin-walled, low-pressure, distensible conduit; even light transducer pressure exceeds intraluminal venous pressure and squeezes the anterior and posterior walls into full apposition — the lumen visually disappears. A vein containing thrombus cannot collapse because the clot is a semisolid mass occupying the lumen; even firm compression leaves a persistent round or oval structure, sometimes with visible clot outline against the vessel wall.
Partial compressibility — where the vein narrows but does not fully coapt — suggests a non-occlusive thrombus adherent to one wall, an important intermediate finding: it confirms clot is present but flow is still passing around it, which has different prognostic and treatment urgency than a fully occlusive thrombus.
The non-compressibility sign alone, on transverse imaging of the common femoral and popliteal veins, has sensitivity and specificity both exceeding 94% for proximal DVT — making 2-point (or 3-point) compression ultrasound at these two sites a validated, rapid point-of-care strategy in emergency settings, without needing to survey the entire venous tree.
While compressibility is the primary diagnostic sign, duplex Doppler adds a functional dimension: color flow imaging shows whether blood is moving through the vein at all, and pulsed-wave spectral Doppler characterizes the shape of that flow over time. Four classic features are assessed at each venous segment — spontaneity, phasicity, augmentation, and competence — and departures from the normal pattern point toward proximal obstruction even when the obstructing clot itself is out of view.
Spontaneity: flow should be detectable without any provocation in the larger proximal veins (femoral, popliteal); its complete absence suggests occlusion somewhere in that segment.
Phasicity: normal venous flow varies rhythmically with respiration — increasing with expiration and decreasing (sometimes ceasing) with inspiration, as diaphragmatic descent raises intra-abdominal pressure and transiently impedes venous return from the legs. This "respirophasic" pattern is best seen in the femoral and common femoral veins. Its loss — a flat, continuous, non-varying signal — implies fixed obstruction proximal to the point of interrogation, most classically from an iliac or IVC thrombus or extrinsic compression.
Augmentation: manual compression of the calf should produce a brisk, transient surge in flow velocity proximal to the compression point, reflecting a patent, unobstructed conduit between the compression site and the probe. A blunted or absent augmentation response localizes an obstruction between the two points.
Competence: with release of calf compression (or Valsalva), flow should cease promptly; more than 1 second of reversed (retrograde) flow indicates valvular incompetence, a common late sequela of prior DVT (post-thrombotic changes) rather than an acute finding.
Directly over an occlusive thrombus, color Doppler typically shows no flow at all — the lumen fails to fill with color despite correct gain and scale settings. Around a non-occlusive thrombus, color flow often reveals a thin rim of flow tracking around the clot, sometimes described as visualizing the thrombus as a filling defect against the color column.
Downstream (distal) to a proximal occlusion, spectral Doppler in the calf veins may show a continuous, monophasic, non-respirophasic pattern because the normal pressure transmission from respiration is dampened by the obstruction upstream. This means Doppler waveform analysis can raise suspicion for thrombus in a segment that was not — or could not be — directly visualized, such as the iliac veins or IVC, which lie too deep or are obscured by bowel gas for direct compression testing.
A monophasic, non-phasic, non-augmenting Doppler signal in an otherwise normally compressible distal vein should always prompt evaluation of the proximal venous system (iliac veins, IVC) — even when direct compression there is technically difficult — because it is an indirect but reliable marker of more central obstruction.
Once thrombus is confirmed, two further characteristics change management: how the clot appears sonographically (which correlates with how recently it formed) and precisely where it sits along the venous tree (which determines pulmonary embolism risk and urgency of treatment). Acute and chronic thrombus look, and behave, quite differently under the transducer.
Acute thrombus (typically <1–2 weeks old) is characteristically hypoechoic to nearly anechoic — it can be easy to miss on grayscale imaging alone, which is exactly why the compressibility test (not visualization of the clot itself) is the primary diagnostic sign. Acute clot tends to distend the vein, making its diameter larger than the adjacent artery, and it is soft and minimally adherent to the wall, sometimes visibly mobile ("free-floating") at its leading edge — a feature associated with higher embolic risk.
As thrombus organizes over weeks to months, fibrin cross-linking, collagen deposition, and partial recanalization increase its echogenicity — chronic thrombus appears bright, heterogeneous, and often retracted, no longer distending the vein and sometimes reduced to an irregular, echogenic band adherent to the wall with visible recanalized channels of flow through or around it. Chronic post-thrombotic changes include a thickened, synechiae-lined vein wall and valve destruction visible as reflux on Doppler.
The deep venous system is conventionally divided at the trifurcation below the knee. Proximal DVT involves the popliteal vein or anything more central — femoral, common femoral, iliac, or IVC. Distal DVT is confined to the calf veins: posterior tibial, peroneal, or (less classically included) the muscular soleal and gastrocnemius veins.
This distinction matters because the calf veins are small-caliber and drain into the much larger popliteal vein; a clot confined there has a comparatively low probability of generating a hemodynamically significant pulmonary embolus. Proximal veins, by contrast, are wide-bore, high-flow conduits directly in line with the IVC and right heart — thrombus here embolizes far more readily and with larger fragment size. Roughly 20–30% of untreated isolated distal DVT will propagate proximally within 1–2 weeks, which is precisely why untreated distal clot is managed with short-interval surveillance rather than simple reassurance.
A clot's clinical weight is set by its address, not just its presence: identical thrombus volume carries dramatically different pulmonary embolism risk depending on whether it sits in a 5 mm posterior tibial vein or a 12 mm common femoral vein.
The final step converts everything gathered — Wells category, compressibility, Doppler pattern, echogenicity, and anatomic location — into a concrete management decision. In modern DVT care, this decision point is fundamentally a fork: anticoagulate, or observe with serial imaging, with a much smaller subset requiring more aggressive intervention such as catheter-directed thrombolysis or IVC filter placement.
1. Estimate pretest probability with the Wells score from history and exam. 2. Low probability (Wells ≤1): order high-sensitivity D-dimer. If negative, DVT is excluded — no imaging required. If positive, proceed to compression ultrasound. 3. Moderate-to-high probability (Wells ≥2): proceed directly to compression ultrasound; D-dimer is not reliable enough to exclude disease in this group. 4. Ultrasound positive (non-compressible segment): characterize location — proximal or distal — and echogenicity (acute vs. chronic) to guide urgency and duration of therapy. 5. Ultrasound negative but pretest probability was moderate-to-high: repeat compression ultrasound in 5–7 days to catch clot that was initially too small, too distal, or technically obscured to detect. 6. Proximal DVT (or symptomatic/extensive distal DVT): begin anticoagulation, typically a direct oral anticoagulant (DOAC), for a minimum of 3 months. 7. Isolated, non-extensive distal DVT in a low-risk patient without severe symptoms: either treat with anticoagulation or pursue serial ultrasound surveillance at ~1 week and ~2 weeks to detect proximal extension, treating only if it occurs.
The proximal/distal split is the dominant variable, but it is not the only one. Even isolated distal DVT is more often treated with anticoagulation when the patient has severe symptoms, risk factors that will persist (active cancer, ongoing immobilization), extensive clot burden involving multiple calf veins, thrombus close to the popliteal vein, or when reliable follow-up for serial scanning cannot be guaranteed. Conversely, a patient with a small, non-extensive, incidentally found distal clot, no ongoing provoking risk factor, and reliable follow-up is a reasonable candidate for surveillance alone, sparing them the bleeding risk of anticoagulation.
For extensive proximal DVT with severe limb-threatening swelling (phlegmasia) or in younger patients with iliofemoral DVT and low bleeding risk, catheter-directed thrombolysis may be considered to reduce post-thrombotic syndrome. An IVC filter is reserved for patients with a confirmed or high-risk DVT/PE who have an absolute contraindication to anticoagulation, not as a routine adjunct.
Bottom line: proximal DVT is anticoagulated, essentially without exception, because of its high embolic potential; isolated distal DVT is a judgment call between anticoagulation and short-interval ultrasound surveillance, weighing bleeding risk, symptom severity, and the reliability of follow-up.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Active malignancy | +1 point | Treatment ongoing, within 6 months, or palliative | Hypercoagulable state |
| Paralysis / immobilization | +1 point | Recent plaster cast of lower extremity | Venous stasis |
| Bedridden ≥3 days / major surgery | +1 point | Within prior 12 weeks, general/regional anesthesia | Stasis + endothelial injury |
| Localized deep-system tenderness | +1 point | Palpation along deep venous course | Direct exam finding |
| Entire leg swollen | +1 point | Diffuse edema of whole limb | Suggests proximal obstruction |
| Calf swelling >3 cm | +1 point | Measured 10 cm below tibial tuberosity, vs. other leg | Objective asymmetry |
| Alternative diagnosis as likely | −2 points | Cellulitis, Baker's cyst, muscle strain | Reduces score, adjusts prior |