You are on page 1of 5

J Thromb Thrombolysis (2014) 37:298302

DOI 10.1007/s11239-013-0945-5

Resident performed two-point compression ultrasound

is inadequate for diagnosis of deep vein thrombosis
in the critically III
Jonathan Caronia Adrian Sarzynski
Babak Tofighi Ramyar Mahdavi Charles Allred
Georgia Panagopoulos Bushra Mina

Published online: 31 May 2013

Springer Science+Business Media New York 2013

Abstract Doppler ultrasonography is a standard in

diagnosis of deep vein thrombosis (DVT) but is often
delayed. Clinician-performed focused vascular sonography
(FVS) has proven to accurately diagnose DVT in the
ambulatory and emergency room settings. Whether trained
medical residents can perform quality FVS in the critically
ill is unknown. Medical residents were trained in a 2-hour
module in FVS assessing for complete compressibility of
common femoral and popliteal veins. Residents imaged
consecutive medical ICU and intermediate care patients
awaiting comprehensive, sonographer-performed and radiologist-interpreted examinations. Sensitivity, specificity,
positive and negative predictive values of the focused
examination were calculated against the comprehensive
study. Fleiss Kappa (j), the degree of agreement between
resident and radiologist, was calculated. Time savings was
measured. Nineteen residents performed 143 studies on 75
patients. Twelve patients had above-the-knee DVTs, a
prevalence of 16 %. All 6 common femoral and 7 of 9
popliteal vein DVTs were identified. None of 6 isolated
superficial femoral DVTs were identified. Sensitivity for
above-the-knee DVT was 63 %, specificity 97 %. Sensitivity for common femoral and popliteal DVT was 86 %,

Presented orally at the American College of Chest Physicians

Meeting in Honolulu, HI, October 2011.
J. Caronia (&)  A. Sarzynski  B. Tofighi  R. Mahdavi 
C. Allred  B. Mina
Division of Pulmonary and Critical Care Medicine, Lenox Hill
Hospital, 100 E 77th Street, New York, NY 10075, USA
G. Panagopoulos
Department of Bio-statistics, Lenox Hill Hospital, 100 E 77th
Street, New York, NY 10075, USA


specificity 97 %. Residents showed substantial agreement

with radiologists for diagnosis of DVT (j = 0.70, SE 0.114,
p \ 0.001).Time from order of a formal ultrasound to a
radiologists read averaged 14.7 h. The two-point compression ultrasound method demonstrated insufficient sensitivity in a cohort of critically ill medical patients due to a
high-incidence of superficial femoral DVT. However, residents demonstrated substantial agreement with radiologists
for the diagnosis of clinically relevant DVT after a 2-hour
course. FVS should include the superficial femoral vein and
is associated with a significant time savings.

Focused vascular sonography  Residents 

Critically ill patients are at high risk for development of
lower extremity deep vein thrombosis (DVT), with a reported incidence of 12 % [1]. Doppler ultrasonography with
graded compression along contiguous venous segments is
widely accepted as a gold standard in diagnosis [2]. Ultrasound studies are traditionally performed by technologists
and interpreted by board-certified radiologists or vascular
specialists. However, delays in performance and interpretation of the comprehensive study are common. Clinicianperformed focused ultrasound evaluation offers the potential
to reduce this delay. Focused vascular sonography (FVS)
performed and interpreted by intensivists and emergency
physicians is becoming accepted practice. The American
College of Chest Physicians (ACCP) recommends that FVS
be included as part of a basic curriculum in intensivist-performed ultrasound [3]. The American College of Emergency
Physicians (ACEP) also endorses performance of FVS by

Diagnosis of deep vein thrombosis

emergency physicians in high risk patients [4]. Training in

FVS to diagnose DVT is currently recommended for all
emergency medicine residents in the United States [5].
The simplified two-point compression method evaluates
the common femoral and popliteal veins for complete
compressibility with 2-D ultrasound and has proven to
accurately detect DVT in the outpatient and emergency
room settings [612]. Crisp et al. [10] reported a sensitivity
of 100 % and specificity of 99 % for clinically-relevant
DVT when the two-point compression method was performed by emergency room attending physicians. Two-point
FVS generally avoids use of color flow and pulsed-wave
Doppler augmentation and relies on the fact that complete
compressibility of a deep vein is the most sensitive and
specific method to diagnose DVT. This simplified method is
often taught at emergency and critical care ultrasound
courses. The original justification for the two-point approach
was provided by Cogo et al. [13], who studied venograms of
562 outpatients with symptoms of DVT, and found that
thrombosis of the superficial femoral vein, a misnomer as it
is a deep vein of the leg, was always associated with more
extensive proximal thrombosis. Some authors have therefore
recommended using a two-point method to exclude DVT in
the emergency room setting [14], while others have included
evaluation of the superficial femoral vein, a three-point
method [15, 16].
Internal medicine residents participate in the care of
critically ill patients at risk for DVT on a daily basis.
Trained internal medicine residents have proven the ability
to perform quality focused sonography, including focused
echocardiography in prior studies [1721]. Echocardiography, which requires imaging an organ in motion in
multiple views, may be more difficult than focused vascular sonography, in which only transverse and longitudinal views of a vein are required. These studies suggest that
medical residents can be trained to perform high quality
FVS for the timely diagnosis of DVT. However, we are the
first to train internal medicine residents in a dedicated
curriculum in focused vascular sonography and to track
their agreement with comprehensive sonography. We also
tracked the time saved by performance of these studies.

Setting and inclusion/exclusion criteria
We performed a prospective, cohort study at Lenox Hill
Hospital, a 652-bed urban, community teaching hospital in
New York City. IRB approval was obtained and all patients
or next-of-kin gave conformed consent to participate.
Consecutive patients admitted to the medical ICU or
intermediate care unit, who were at least 18 years old, and


had clinical symptoms suggestive of lower extremity DVT

as determined by the intensivist attending physician were
eligible for inclusion. Patients with a known history of
DVT in either lower extremity or in active hemodynamic
decompensation were excluded. If patients had femoral
central lines present, the femoral area was not imaged by
residents but the popliteal area was imaged ipsilaterally.
Residents also gave informed consent to participate.
Resident training and focused examination
Residents on a two or four-week elective in critical care
ultrasound were trained in a standardized 2-hour course,
which included didactic and practical instruction using a
SonositeTM M-Turbo (Sonosite, Inc, Bothell, WA, USA) or
ZonareTM Z-One-Ultra (Zonare Medical Systems, Mountain View, CA, USA) ultrasound unit. Residents were
trained in ultrasound physics and proper use of the machine
as well as patient safety including the ALARA (as low as
reasonably achievable) principle. Residents were trained by
a board-certified critical care attending and a full-time
research fellow, both of whom had received formal training
in point-of-care ultrasound. Residents were taught the
validated 2 point-compression method of FVS originally
suggested by Cogo (Arch Intern Med, 1993) to identify the
common femoral vein, sapherofemoral junction, and popliteal veins in transverse view with a linear 7.5 MHz probe.
The superficial femoral vein was not evaluated, nor was
any structure distal to the popliteal vein. A focused study
was considered to be positive if the resident was unable to
completely compress the common femoral vein or the
popliteal vein with graded compression. Color and pulsed
wave Doppler were not used in the diagnosis of DVT;
however residents were allowed to use these modalities for
anatomical distinction between vein and artery. Loss of
complete compressibility was the only criteria for a positive study. No clinical decisions were made based on resident FVS.
Comparison with formal study
Trained ultrasound technicians performed comprehensive
lower extremity Doppler studies within 24 h of resident
FVS. The formal studies were then read by board-certified
radiologists who were unaware of the resident findings.
The resident was blinded to the formal read and to any
impressions of the technologist, if the formal study was
performed before the resident study. The time between the
initial order and the radiologists reading was calculated
based on electronic medical records. The order time is
logged on the electronic order system, and the radiology
report is time stamped when read by a radiologist. Residents also recorded the time necessary to perform a



J. Caronia et al.

unilateral examination. After residents submitted their

focused studies to the research fellow (JC) they reviewed
their image acquisition and the comprehensive study with
the research fellow, however they were not able to amend
their answers based on this review.

Table 2 Sensitivity, specificity, positive and negative predictive

values of the resident focused examinations with the comprehensive
study as reference, by anatomical location

Above the
knee DVT
included) (%)

DVT (%)

DVT (%)
















Statistical analysis
Sensitivity, specificity, positive predictive value, and negative predictive value of resident exams were calculated using
the formal examination as the reference standard. Fleiss
Kappa, the degree of inter-observer agreement between
residents and radiologists for the diagnosis of DVT, was our
primary outcome. The calculation was performed using
standard software (SPSS version 20.0, IBM software).

DVT (%)

Nineteen residents performed 143 studies on 75 patients
from December 2010 to August 2011, with a study defined
as a unilateral sonographic examination. Eight patients
declined to participate. Each resident performed an average
of 8 studies (SD 5). The maximal number of examinations
performed by a resident was 23 and the minimum was 2.
The mode number of examinations was 8.5. Mean patient
age was 70 years old. Thirty-eight female and 37 male
patients participated in the study. Formal studies were read
by one of 7 board-certified radiologists.
Twelve of 75 patients had above-the-knee DVTs on the
formal study, a prevalence of 16 %. Residents correctly
identified 13 of 21 DVTs in the cohort. Anatomical distribution on the formal study of DVTs is shown in Table 1.
All 6 common femoral and 7 of the 9 popliteal vein DVTs
were correctly identified by the residents. None of 6 isolated superficial femoral DVTs were identified by the residents. The sensitivity and specificity of the studies based
on anatomical distribution are listed in Table 2. Residents
showed substantial inter-observer agreement with radiologists for the diagnosis of above-the-knee DVT, demonstrating a Fleiss Kappa = 0.70 (SE 0.114, p \ 0.001).The
average time between the order of a formal study to its read
by a radiologist was 14.7 h. The average unilateral twopoint examination took approximately 5 min.

Table 1 Anatomic distribution of DVTs identified on comprehensive

Anatomic location

Number of DVTs

Common femoral vein

Superficial femoral vein

Popliteal vein


Resident FVS failed to demonstrate sufficient sensitivity in

a cohort of critically ill patients due to a high incidence of
superficial femoral vein thrombosis. However, after a
2-hour course in FVS based on ACCP recommendations,
residents showed substantial agreement with radiologists
for the diagnosis of clinically-relevant DVT. Location of
lower extremity DVT in the critically ill, versus other
patient populations, has not been well studied. We found
only one publication, an abstract, addressing this issue.
Zytaruk et al. [22] found that critically ill medical and
surgical patients were likely to have DVT at multiple
venous segments. Furthermore, the prevalence of superficial femoral DVT in Zytaruks cohort was nearly 40 %; it
was the most common site of DVT in the critically ill.
Sixteen percent of DVT were catheter-related. Perhaps
instrumentation of the femoral vessels, common in the
critically ill, is a risk factor for development of superficial
femoral vein DVT. The objective of our study was to train
the resident to identify DVT at the common femoral vein,
sapheno-femoral junction, and popliteal veins, where they
demonstrated a sensitivity of 86 % and specificity of 97 %.
Some may ask why we ignored imaging of the superficial
femoral vein. The answer is that the two-point method,
which does not include the superficial femoral vein, has
been well-validated in emergency room patients. Crisp [10]
used the same focused two-point method performed by
emergency attendings in emergency room patients and
reported a sensitivity of 100 % and specificity of 99 % for
above- the-knee DVT. Frazee [9] also used the two-point
compression method in a cohort of 76 emergency room
patients by emergency physicians and demonstrated a
sensitivity of 88.9 % and specificity of 75.9 % when FVS
was compared against comprehensive vascular sonography.
Blavias [11] reported a kappa coefficient of 0.90 between

Diagnosis of deep vein thrombosis

emergency physicians and radiologists for all clinically

relevant DVT training the emergency physicians in the
same two-point method that we used. Our work suggests
that the two-point method should not be used in critically
ill patients. We are not the first to dispute Cogos findings.
Maki [23] demonstrated that up to 22 % of lower extremity
DVTs could be isolated to the superficial femoral vein.
Frederick [24] showed DVTs were limited to the superficial
femoral vein without proximal extension in 4.6 % of cases
of acute DVT and thus would be missed with an abbreviated two-point method. In our cohort, the prevalence of
superficial femoral vein DVT was 8 %, and the exclusion
of imaging the superficial femoral vein was therefore significant. In a recent study, critical care attendings and
fellows demonstrated a sensitivity of 86 % and specificity
of 96 % for DVT using a more comprehensive method that
evaluated the superficial femoral vein [23]. Addition of the
superficial femoral vein to the femoral and popliteal areas
could substantially increase the sensitivity for relevant
DVT without significantly increasing the time required for
the study.
The common femoral vein was the largest vessel evaluated and residents achieved a sensitivity of 100 % and a
specificity of 97 % at this region. The somewhat lower
sensitivity and specificity at the popliteal vein may be
explained by the smaller size of the vessel and the technical
difficulty of imaging the popliteal vein compared to the
common femoral vein. Positioning of the patient often
requires bending the knee and externally rotating at the hip
which may be difficult in critically ill patients. Residents
may need more than 2 h of initial training to accurately
image and diagnose DVT at the anatomically smaller
popliteal area. We found a time delay of 14.7 h between
the order of the comprehensive ultrasound and its interpretation by a radiologist. Kory et al. [25] reported a time
lag of 13.8 h. These are significant wait times in a critically
ill patient, and argues for implementation of FVS at the
fellow and resident levels.
Limitations of our study include a relatively small
number of DVTs and significant variation in the number of
examinations performed by the residents. This was a single
center study. These factors may limit the generalizability of
our results. A similar study that includes compressibility of
the superficial femoral vein should be repeated with larger
numbers in multiple centers to further assess generalizability of the results. Our study was not powered to detect
the number of studies necessary to become proficient, and
this is an important area to be addressed. Furthermore,
focused sonography has never been shown to improve
outcomes such as mortality or ICU length of stay.
This paper is one of the first to describe the accuracy of
focused vascular sonography performed by internal medicine residents in the critically ill. Our data suggest that the


two-point compression method should be modified to a

three-point method that includes the superficial femoral
vein. This three-point method is recommended by the
ACEP [26]. The limitations of focused sonography should
be understood and respected, including the lower sensitivity for the smaller popliteal vein. We believe that
internal medicine residents can be trained to diagnose DVT
in the critically ill at bedside with a three-point method that
includes the superficial femoral vein.
Acknowledgments This work was supported by an educational
grant from the New York State Department of Health, ECRIP, Empire
Clinical Research Investigator Program.
Conflict of interest

No authors have any other relevant conflicts of

1. Marik PE, Andrews L, Maini B (1997) The incidence of deep
venous thrombosis in ICU patients. Chest 111:661664
2. Qaseem A, Snow V, Barry P et al (2007) Current diagnosis of
venous thromboembolism in primary care: a clinical practice
guideline from the American Academy of Family Physicians and
the American College of Physicians. Ann Intern Med 146:454459
3. Mayo PH, Beaulieu Y, Doelken P et al (2009) American college
of chest physicians/la societe de reanimation de langue francaise
statement on competence in critical care ultrasonography. Chest
4. Rios M, Lewis R, Saul T (2009) Focus on: emergency ultrasound
for deep vein thrombosis. Accessed 5 Jan 2012
5. Akhtar S, Theodoro D, Gaspari R, Tayal V, Sierzenski P, Lamantia J, Stahmer S, Raio C (2009) Resident training in emergency ultrasound: consensus recommendations from the 2008
council of emergency medicine residency directors conference.
Acad Emerg Med 16(Suppl 2):3236
6. Bernardi E, Camporese G, Buller HR et al (2008) Serial 2-point
ultrasonography plus D-dimer vs. whole-leg color-coded Doppler
ultrasonography for diagnosing suspected symptomatic deep vein
thrombosis: a randomized controlled trial. JAMA 300:16531659
7. Cogo A, Lensig AW, Koopman MM et al (1998) Compression
ultrasonography for diagnostic management of patients with
clinically suspected deep vein thrombosis: prospective cohort
study. BMJ 316:1720
8. Lensing AW, Prandoni P, Brandjes D et al (1989) Detection of
deep-vein thrombosis by real-time B-mode ultrasonography.
N Engl J Med 320:342345
9. Frazee BW, Snoey ER, Levitt A (2001) Emergency department
compression ultrasound to diagnose proximal deep vein thrombosis. J Emerg Med 20:107112
10. Crisp JG, Lovato LM, Jang TB (2010) Compression ultrasonography of the lower extremity with portable vascular ultrasonography can accurately detect deep venous thrombosis in the
emergency department. Ann Emerg Med 56:601610
11. Blaivas M, Lambert MJ, Harwood RA, Wood JP, Konicki J
(2000) Lower-extremity Doppler for deep venous thrombosiscan
emergency physicians be accurate and fast? Acad Emerg Med
12. Heijboer H, Buller HR, Lensing AW, Turpie AG, Colly LP, Ten
Cate JW (1993) A comparison of real-time compression ultrasonography with impedance plethysmography for the diagnosis
of deep vein thrombosis in symptomatic outpatients. N Engl J
Med 329:13651369


13. Cogo A, Lensing AW, Prandoni P, Hirsh J (1993) Distribution of
thrombosis in patients with symptomatic deep vein thrombosis.
Implications for simplifying the diagnostic process with compression ultrasound. Arch Intern Med 153:27772780
14. Fox JC, Bertoglio KC (2011) Emergency physician performed
ultrasound for DVT evaluation. Thrombosis. doi:10.1155/2011/
15. Trottier SJ, Todi S, Veremakis C (1996) Validation of an inexpensive B-mode ultrasound device for detection of deep vein
thrombosis. Chest 110:15471550
16. Magazzini S, Vanni S, Toccafondi S et al (2007) Duplex ultrasound in the emergency department for the diagnostic management of clinically suspected deep vein thrombosis. Acad Emerg
Med 14:216220
17. Bailey BP, Ault M, Greengold NL, Rosendahl T, Cossman D
(2001) Ultrasonography performed by primary care residents for
abdominal aortic aneurysm screening. J Gen Int Med 16:845849
18. Vignon P, Dugard A, Abraham J et al (2007) Focused training for
goal-oriented hand-held echocardiography performed by non
cardiologist residents in the intensive care unit. Intensive Care
Med 33:17951799
19. Croft LB, Duvall WL, Goldman ME (2006) A pilot study of the
clinical impact of hand-carried cardiac ultrasound unit in the
medical clinic. Echocardiography 23:439446
20. Hellman DB, Whiting-OKeefe Q, Shapiro EP, Martin LD,
Martire C, Ziegelstein RC (2005) The rate at which residents


J. Caronia et al.







learn to use hand-held echocardiography at the bedside. Am J

Med 118:10101018
Chalumeau-Lemoine L, Baudel JL, Das V, Arrive L et al (2009)
Results of short-term training of nave physicians in focused
general ultrasonography in an intensive-care unit. Intensive Care
Med 35:17671771
Zytaruk N, Heels-Ansdell D, McDonald J et al (2011) Venous
segments affected in critically ill patients with deep vein
thrombosis. Am J Respir Crit Care Med 183:A2382
Maki DD, Kumar N, Nguyen B, Langer JE, Miller WT Jr, Gefter
WB (2000) Distribution of thrombi in acute lower extremity deep
venous thrombosis: implications for sonography and CT and MR
venography. Am J Radiol 175:12991301
Frederick MG, Hertzberg BS, Kliewer MA, Paulson EK, Bowie
JD, Lalouche KJ, DeLong DM, Carroll BA (1996) Can the US
examination for lower extremity deep venous thrombosis be
abbreviated? A prospective study of 755 examinations. Radiology 199:4547
Kory PD, Pellecchia CM, Shiloh AL, Mayo PH, DiBello C,
Koenig S (2011) Accuracy of ultrasonography performed by
critical care physicians for the diagnosis of DVT. Chest
American College of Emergency Physicians (2006) Emergency
ultrasound imaging criteria compendium. Ann Emerg Med.