Curated insights • How it Works • Practical Pearls • Evidence Base
Validated in adult outpatients and ED patients with suspected DVT. Not validated in pregnancy, patients on therapeutic anticoagulation, or those with prior limb bypass surgery. Evidence for inpatient use is mixed — one study reported a 5.9% failure rate in 1135 inpatients, suggesting the score should be used cautiously in hospitalised patients.
The Wells DVT score assigns 1 point for each of 9 clinical features associated with DVT, then subtracts 2 points if an alternative diagnosis is considered at least as likely. Net score ranges from −2 to +9. The original 1997 model used 9 items; the 2003 modification added a 10th (previously documented DVT, +1 point) and introduced the binary "DVT likely / DVT unlikely" classification. Both the 3-tier and 2-tier systems are validated and in clinical use.
"DVT Unlikely" = score < 2 (prevalence ~6%); "DVT Likely" = score ≥ 2 (prevalence ~28%). This binary classification is used in most UK, European, and ACCP-aligned guidelines. In the DVT Unlikely group, a negative high-sensitivity D-dimer carries a failure rate of approximately 1.2% across 13 validated studies — within accepted safety thresholds.
The "alternative diagnosis at least as likely" criterion (−2 points) is the most contested element of the score. Interobserver reliability for this item alone is only moderate. The overall Wells score does, however, achieve excellent interobserver agreement (κ ≈ 0.85) because the other nine items are objective. Clinicians with less experience or risk-averse dispositions tend to under-apply this deduction, inflating the proportion scored as DVT Likely.
In patients older than 50 years, apply the age-adjusted D-dimer threshold: age × 10 µg/L (fibrinogen equivalent units). This threshold safely increases specificity without meaningfully sacrificing sensitivity, reducing unnecessary ultrasound referrals by approximately 20% in patients over 75. Do not apply age-adjustment in DVT Likely patients — proceed to imaging.
Not all D-dimer assays are equivalent. ELISA and quantitative immunoturbidimetric assays (e.g., STA-Liatest D-Di Plus, Vidas D-Dimer Exclusion) have sufficient sensitivity (>95%) to safely exclude DVT in low- or unlikely-probability patients. Latex agglutination assays and SimpliRED whole-blood assays have lower sensitivity and are not recommended for rule-out without concurrent clinical probability assessment. Always confirm which assay your institution uses.
If the initial compression ultrasound is negative in a high-probability patient (score ≥ 3), do not discharge without a plan for repeat imaging. Up to 2% of high-probability patients with an initially negative proximal compression ultrasound will have DVT confirmed on repeat scanning at 7 days. Whole-leg ultrasound (including calf veins) may reduce but does not eliminate this gap. For patients where repeat scanning is not feasible, consider same-day venography.
When DVT symptoms occur alongside suspected PE, the Wells PE Score should be applied concurrently. In outpatients with a Wells PE score < 2, the PERC rule can be applied to identify a subset where no further testing (including D-dimer) is warranted. PERC criteria include: age < 50, pulse < 100, SpO2 ≥ 95%, no haemoptysis, no oestrogen use, no prior DVT/PE, no unilateral leg swelling, no recent hospitalisation or surgery. All 8 must be negative.
A 2024 Norwegian study (Halstensen et al., n = 1312) derived and internally validated a 2-variable objective score using only tenderness along deep veins and prior VTE history. This "derived DVT score" achieved NPV of 99.4% with D-dimer, comparable to Wells, with a failure rate of 1.8% vs 1.5% for Wells — though at the cost of 14% more patients requiring further workup. External validation is pending and clinical adoption is not yet warranted.
Once DVT is confirmed, anticoagulation should be initiated without delay. DOACs (rivaroxaban, apixaban) are first-line for most patients. LMWH bridging to warfarin remains appropriate in cancer-associated DVT (where LMWH or edoxaban/rivaroxaban are preferred) and in selected patients with renal impairment. Duration: 3 months minimum; reassess for extended therapy based on provoked vs unprovoked status, recurrence risk, and bleeding risk.
Value of assessment of pretest probability of deep-vein thrombosis in clinical management.
Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis.
Does this patient have deep vein thrombosis?
Exclusion of deep vein thrombosis using the Wells rule in clinically important subgroups: individual patient data meta-analysis.
Wells criteria for DVT is a reliable clinical tool to assess the risk of deep venous thrombosis in trauma patients.
Development and internal validation of a simple clinical score for the estimation of the probability of deep vein thrombosis in outpatient emergency department patients.
A Canadian haematologist at the University of Ottawa who developed both the DVT and PE clinical prediction rules throughout the 1990s and early 2000s. His work transformed the management of thromboembolism by demonstrating that structured clinical scoring could safely replace reflexive imaging, reducing costs, radiation exposure, and unnecessary anticoagulation across millions of annual patient encounters.
Prior to the Wells score, DVT diagnosis relied heavily on contrast venography — the historical gold standard. This was invasive, required contrast injection through a foot vein, and was not widely accessible. Impedance plethysmography and radiolabelled fibrinogen uptake testing were also in use but lacked specificity. The Wells criteria, published in 1997, provided the first clinically usable framework for integrating history and examination findings into a structured pretest probability — enabling safe use of non-invasive D-dimer assays as the first-line test and reserving ultrasound for patients who truly needed it.
Curated insights • How it Works • Practical Pearls • Evidence Base
HEART score alone is not sufficient. Always combine with two serial high-sensitivity troponin measurements. A HEART 0–3 + two negative hs-troponins = 30-day MACE risk < 2% → safe early discharge.
LBBB pattern (paced or Sgarbossa-negative) can score a false +2 on ECG domain. In LBBB patients, apply Sgarbossa or modified Sgarbossa criteria to assess for STEMI equivalence separately.
Chest pain in the emergency room: value of the HEART score.
A prospective validation of the HEART score for chest pain patients at the emergency department.
Developed in the Netherlands by AJ Six and colleagues in 2008 as a simple bedside tool combining five clinical domains, each scored 0–2. Its elegant acronym (History, ECG, Age, Risk factors, Troponin) allows recall without a calculator. It has since become the dominant chest pain pathway tool in emergency medicine globally.
