Ultrasound G.E.L. – Accuracy of 3 Point DVT Exam

Originally published on Ultrasound G.E.L. on 2/26/18 – Visit HERE to listen to accompanying PODCAST! Reposted with permission.

Follow Dr. Michael Prats, MD (@PratsEM) from Ultrasound G.E.L. team!

Comparison of the Accuracy of Emergency Department-Performed Point-of-Care-Ultrasound (POCUS) in the Diagnosis of Lower-Extremity Deep Vein Thrombosis

Journal of Emergency Medicine Jan 2018 – Pubmed Link

Take Home Points

1. Despite adequate training in POCUS, the three point examination is lacking in both sensitivity and specificity in this population.

2. Decision making based on point-of-care DVT exams should take into account the pre-test probability of DVT, the experience of the operator, and should likely include close follow up.


Deep Vein Thrombosis (DVT) is a potentially dangerous pathology that is typically diagnosed by a comprehensive Doppler ultrasound performed by a sonographer. Unfortunately, comprehensive ultrasonography is not always readily available, and patients may be required to return for testing after discharge. This system can result in patients lost to follow up, delayed diagnosis, and unnecessary treatment of patients with blood thinning medications. Previous research has shown that point of care ultrasound can be fairly accurate for this diagnosis, although studies have been mixed. One criticism has been that prior point of care protocols were two-point compression protocols, examining the common femoral and the popliteal vein but not examining the femoral vein. This might lead to a decreased sensitivity and missed DVTs. The current study examined whether emergency physicians could accurately diagnose DVTs, using a point of care, three point compression protocol.


How accurately can emergency physicians diagnose lower extremity DVTs using a point of care ultrasound (POCUS), three point compression protocol?


The patient population consisted of a convenience sample of patients over the age of 18 with a suspected DVT who presented to a regional hospital’s Emergency Department in Andalucía, Spain

Between March 2012 and May 2014.


  • ≥18 years old
  • Signs or symptoms concerning for DVT


  • Diagnosed DVT
  • DVT in past 6 months
  • Doppler US in past 30 days
  • Low risk by Wells AND negative D-dimer
  • Lost to follow up
  • ED physician trained in US was not available


This was a prospective cross-sectional study convenience sample (only when US trained people, 8am-midnight during the week) and diagnostic test assessment.

Patients with a suspected DVT had a Well’s score calculated and had a D-dimer test drawn.

If low risk by Well’s score and a negative D-dimer, they were excluded from study analysis.

All other patients underwent a POCUS exam by the Emergency Physician.

Each patient then had a comprehensive doppler ultrasound exam, performed by an ultrasonographer and interpreted by a radiologist.

The results between the POCUS and comprehensive exams were then analyzed and compared. For the purposes of analysis, an indeterminant POCUS was considered positive.

Follow up by telephone, software, and DVT clinic

  • 3, 5, and 12 months after first visit

The exams were timed from when the probe was placed on the patient to the time when the study was completed.


The article provides an algorithm for enrolled patients.

  1. Everyone with suspected DVT gets Modified Wells score and serum D-dimer
  2. If Wells score < 2 and D-dimer negative → excluded from study, no further testing obtained
  3. If Wells score < 2 and D-dimer positive → they get a POC ultrasound, if positive → Treatment + Follow up comprehensive doppler ultrasound within 48 hours.
  4. If POC ultrasound negative → No treatment + Follow up comprehensive doppler ultrasound within 48 hours.
  5. If Moderate to High Risk based on Wells → these patients all received D-dimers and POCUS but according to the algorithm, they were ALL discharged with treatment and follow up for comprehensive doppler ultrasound within 48 hours.

Some notes on this algorithm:

  • There is no possibility for admission
  • All patients with Wells score > 2, were treated for DVT regardless of d-dimer and POCUS.

Who did the ultrasounds?

The study population was 11 Emergency Physicians who each had at least 5 years of EM practice after training.

They took part in 3 days US training course that included 4 hours of DVT US instruction.

They then spent 6 months performing the exams in the Emergency Department with interspersed training sessions with radiology.

All had completed >30 POCUS exams for DVT before enrolling patients.

The Scan

Linear probe used


Study Protocol Description:

  • The femoral vein of the affected limb was scanned from the junction of the great saphenous vein distally to the junction of the common femoral, superficial femoral, and deep femoral veins. The probe was then applied to the popliteal fossa, the popliteal vein was identified and scanned 2 cm distal to the popliteal vein and the proximal aspects of its trifurcation into the anterior tibial vein, the posterior tibial vein, and the peroneal vein.
  • The study was considered positive for DVT if any vein was not compressible or if a thrombus was visualized.

5 Minute Sono – How to Scan for DVT

Free Ibook on POCUS for DVT – all you need to know!


The Patients

N = 109

  • 84% had positive d-dimer
  • 88% mod/high risk on Wells score
  • 17% hospital admission (not sure how that happened with their algorithm)
  • 11% had complications such as pulmonary embolism, post-thrombotic syndrome, and chronic thromboembolic pulmonary hypertension

Primary Outcomes

Three-point compression POCUS for lower limb DVT

Sensitivity 93.2% (CI 83.8-97.3%)

Specificity 90.0% (CI 78.6-95.7%)

Accuracy 91.7% (CI 85-95.6%)

+LR 9.32 (CI 4.05-21.47)

-LR 0.08 (CI 0.03-0.19)

Other Findings

ED Ultrasound compared to radiology duplex ultrasound: Kappa = 0.83

Median Time to Completion: 5 minutes, 4 seconds (IQR 6 min 14 s – 11 min 27 s). Max time 20 min 6 s.

False Negatives (4)

  • All were cases of popliteal thrombosis
  • Three of these were in obese patients with large superficial veins in the popliteal fossa that were compressible and thought to be popliteal veins. Radiology confirmed DVT.
  • Single case of with distended collateral veins that were erroneously thought to be deep veins.
  • 3/4 of these patients had positive d-dimer and were treated with anticoagulation.

False Positives (5)

  • 2 cases of lymph node thought to be DVT
  • 1 case of ruptured Baker’s cyst
  • 1 case of superficial thrombophlebitis
  • 1 case of patient admitted with pulmonary embolism. ED scan showed DVT, comprehensive 48 hours later was negative.

False studies were done in patients who tended to have higher BMI compared to those with true studies (34.7 vs 28.5 kg/m2; p < 0.001).


Dedicated training for Emergency Physicians. These physicians probably had more training than some of these other studies.

Confirmatory test occurred up to 48 hours after the initial test, which may affect results. Since treatment was started in many of these patients, who knows if the exams may have changed.

The examiners didn’t examine calf DVTs so this data will not apply to that population.

This only applies to the population with moderate to high risk for DVT. Admirably, the authors did not want to inflate the test characteristics by including many with very low risk for DVT.

Gold standard (radiology duplex) could have missed DVTs as this is not perfect. However, there was pretty good follow up with outpatient appointments and telephone calls.


This study had a lower characteristics than previously reported. In this 2013 metanalysis, most of the studies had better sensitivity and specificity. It is understandable that the sensitivity would be low because the point of care exams are not as thorough as comprehensive sonographer performed exams. What is more concerning is that lack of specificity in the current study. Many practitioners of POCUS are probably used to trusting the specificity of POCUS to rule in the diagnosis of DVT when there is a non-compressible vein on the exam. This study would discourage that with a specificity of only 90%. The problem is that with the lower numbers, even 5 false positives can drop the accuracy significantly.

As always, it is a game of pre-test probability + test = post-test probability. Although they were excluded in this study, if you have a very low pre-test probability, even this sub-par test may help you rule out the disease. Likewise, if you have a very high pre-test probability, this test may help you to rule it in. Obviously, you are at risk for confirmation bias here, but nonetheless, there is potential for this test to change your management. This is something only you can decide how to safely incorporate into your practice.

This study adds to the others on POCUS for DVT that are all over the map in terms of accuracy. It was well performed study, but did have limitations that could be used to justify not believing these results. My guess is that people who thought POCUS was inaccurate will site this as evidence to continue that belief, whereas people who think POCUS is accurate will think they can personally outperform the accuracy seen here. Who do you think is closer to the truth?

Take Home Points

1. Despite adequate training in POCUS, the three point examination is lacking in both sensitivity and specificity in this population.

2. Decision making based on point-of-care DVT exams should take into account the pre-test probability of DVT, the experience of the operator, and should likely include close follow up.

Our score

3 Probes

Cite this post as

Chris San Miguel, MD. Accuracy of 3 Point DVT Exam. Ultrasound G.E.L. Podcast Blog. Published on February 26, 2018. Accessed on May 21, 2021. Available at https://www.ultrasoundgel.org/39.

Leave a Reply

Your email address will not be published. Required fields are marked *