DEEP VEIN THROMBOSIS

See also Pulmonary Embolus and Prophylaxis for DVT & PE

The discussion below applies to lower-extremity DVT; upper-extremity DVT occurs but is less common.

I. Risk factors:

  1. Age > 60yo
  2. Venous stasis
  3. Trauma
  4. Recent surgery
  5. Immobility
  6. Hypercoagulable States: Ca (pancreas, ovary, liver, brain); Pregnancy; Polycythemia; Dehydration; Estrogens (OCP's or HRT); Inflammatory Bowel Disease
  1. Note that despite the association with Ca, it imay not be necessary do extensive w/u for occult malignancy in pts with idiopathic DVT.
    1. Series of 142 pts with DVT and no known risk factor; 12% had Ca, but all of those had at least one abnormal finding on Hx (e.g. weight loss), Px, lab tests (e.g. anemia), or CXR (Ann. Int. Med. 125:785, 1996-JW)
    2. Strongest association between DVT or PE and Ca was in pts < 60yo in a series of 26,653 pts (NEJM 338:1169, 1998--AFP)

II. Clinical presentation:

  1. Lower leg swelling
  2. Pain (dull, aching) + tenderness
  3. Heat
  4. Non-pitting edema
    1. Presence of edema with varicose veins suggests DVT as cause of varicosities: don't operate!

III. Diagnosis

  1. Ultrasound
    1. Simple compression ultrasonography
      1. Highly sensitive for proximal DVT but less sensitive for DVT in distal calf veins
      2. If used as the diagnostic modality for DVT, it's generally recommended that the study be repeated at 5-7d to look for proximal thrombus propagation
    2. "Duplex" ultrasonography (compression ultrasonography + Doppler ultrasonography)
      1. Sensitivity for DVT in the calf veins (as opposed to the "proximal" veins, i.e. popliteal, femoral, and proximal to those) can be limited; some advise repeat u/s 5-7d after negative result if exam was limited.
      2. For LE DVT--Negative duplex u/s of LE was associated with NPV of 94.1% for clinical signs/sx of DVT over 3mos after study done in a cohort of 375 pts with suspected DVT who received no anticoagulation (Ann. Int. Med. 140:1052, 2004--AFP)
      3. In a cohort study in 526 outpatients with suspected LE DVT, in which all pts with negative u/s (visualizing common and superficial femoral veins, popliteal veins, and all 3 calf-vein sets) were followed x 3mos without warfarin, only one was dx'd with venous thromboembolism during the f/u period. (Radiology 237:348, 2005--JW).
        For UE DVT--In a prospective study of 102 pts with suspected UE DVT, all of whom underwent doppler u/s & contrast venography (the latter used as a gold standard), the former was ass'd with sensitivity of 82% and specificity also of 82% (Ann. Int. Med. 136:865, 2002--JW)
  2. Impedance plethysmography
  3. Radionucleide phlebography
  4. Venogram
  5. Magnetic Resonence Direct Thrombus Imaging
    1. Ass'd with sensitivity of 95% and specificity of 91% for DVT (using venography as the gold standard) in a trial of 104 pts with suspected DVT (Ann. Int. Med. 136:89, 2002--JW)
  6. Clinical model for diagnosis (JAMA 279:1094, 1998--AFP)
    1. Point system as follows; "low pretest prob." if < 1; "mod." if 1-2; "high" if > 2
      1. Active cancer (tx ongoing < 6mos ago or palliative)--1 point
      2. Paralysis, paresis, or recent LE cast--1 point
      3. Bedridden for > 3d or major surgery in < 4wks--1 point
      4. Localized tenderness along dist. of deep venous system--1 point
      5. Swelling of entire leg--1 point
      6. Calf swelling by > 3cm c/w asymptomatic leg, measured 10cm below tibial tuberosity--1 point
      7. Pitting edema, greater in the symptomatic leg if bilateral--1 point
      8. Collageral superficial veins (nonvaricose)--1 point
      9. Alternative dx at least as likely as DVT--Minus 2 points
    2. Low pretest prob: if u/s normal, rules out dx; if abnormal, go to venography
    3. Moderate pretest prob: if u/s abnormal, tx as DVT; if normal, no tx but repeat u/s in 1wk and take the result as gospel
    4. High pretest probability: if u/s abnormal, tx as DVT; if normal, go to venography
  7. Plasma D-dimer--See also D-Dimer for Diagnosis of PE
    1. In a series of 686 pts with suspected DVT and negative initial ultrasound, those with normal plasma D-dimer had incidence of 0.2% thromboembolic complication during f/u period (summary didn't say how long), as opposed to 8% for those with abnormal D-dimer (BMJ 317:1037, 1998--JW)
    2. Negative predictive value for DVT was sig. lower in pts with Ca & suspected DVT than those w/o known Ca & suspected DVT (79% vs. 97%; Ann. Int. Med. 131:417, 1999--JW)
    3. In a series of 445 pts with suspected DVT, low clinical suspicion (based on 9-item checklist) plus negative D-Dimer assay ass'd with 3mo incidence of 1/177 developed symptomatic DVT (Ann. Int. Med 135:108, 2001--JW)
    4. In a study of 704 pts with suspected DVT, positive plasma D-dimer (of any value) had 99% sensitivity for DVT on duplex ultrasound (on either intake or at 1wk; all pts had at both times unless first showed DVT) in pts not on anticoagulants but a sensitivity of only 75% for patients on anticoagulants (Am. J. Med. 112:617, 2002--AFP)
    5. In a trial of 896 pts presenting with suspected LE DVT but low clinical suspicion randomized to (D-dimer testing followed by LE duplex ultrasound if D-dimer abnormal) vs. ultrasound alone, there was no sig. diff. in incidence of confirmed venous thromboembolic events over 3mos of f/u (NEJM 349:1227, 2003--abst)
    6. In a systematic review of 12 studies evaluating the combination of plasma D-dimer and estimations of pre-test probability based on a point scale for clinical findings, with confirmation by another diagnostic measure (u/s, venography, or plethysmography) and at least 3mo f/u, the 3mo incidence of DVT was 0.4% in pts with a normal D-dimer and low-medium pretest probability (BMJ 329:821, 2004--JW)

IV. Prevention

  1. See Prophylaxis for DVT & PE
  2. Also, HMG CoA-reductase Inhibitors (aka "statins") may be associated with reduction in risk of venous thromboembolism
    1. In a study in 17,802 healthy adults with LDL < 130 and elevated C-reactive protein randomized to rosuvastatin 20mg/d vs. placebo, over median 1.9y f/u, rosuvastatin recipients had sig. lower incidence of symptomatic venous thromboembolism (0.18 vs. 0.32 events per 100 person-years) ("JUPITER" trial; NEJM 360:1851, 2009-JW)

V. Treatment

  1. Anticoagulation
  1. One of the following until stable on oral warfarin (i.e. until  prothrombin time INR is therapeutic, typically around 5 days):
  1. Adjusted dose IV unfractionated Heparin
  2. Fixed-dose SQ unfractionated Heparin
    1. In a study in 697 ps with acute venous thromboembolism randomized to fixed-dose subcutaneous unfractionated heparin (333 U/kg x 1 then 250 U/kg BID; no PTT monitoring) vs. low-molecular-weight heparin (dalteparin or enoxaparin 100 IU/kg), all of whom got oral warfarin x at least 3mos, there were no sig. differences over initial 10d incidence of major bleeding or recurrent venous thromboembolism over 3mos (JAMA 296:935, 2006--JW)
  3. Low-Molecular-Weight Heparin
    1. In a meta-analysis of 11 randomized trials, LMWH was ass'd with sig.lower mortality at 3-6mos than with unfractionated heparin (5% vs. 7%); also sig. lower rate of major bleeding but this did not entirely account for the mortality benefit (Ann. Int. Med. 130:800, 1999--JW)
    2. In 900 pts with symptomatic acute DVT randomized to Heparin IV (adjusted-dose, target PTT 55-80sec,) vs. Enoxaparin SQ (1.5mg/kg QD or 1mg/kg BID), either until therapeutic INR on Warfarin which was continued x 3mos, 3mo recurrence rates, major hemorrhage incidence, death rate, or PE incidence were not sig. diff. for enoxaparin QD vs. BID vs. Heparin; 32% of pts had PE (Ann. Int. Med. 134:191, 2001--JW)
    3. In a study of 2,205 pts with acute symptomatic DVT randomized to fondaparinux vs. enoxaparin (given until INR > 2.0 on warfarin; at least 5d), 3mo incidence of symptomatic recurrent venous thromboembolism was not sig. diff. in the two groups; neither was major bleeding (Ann. Int. Med. 140:867, 2004--abst)
  1. Warfarin (adjusted-dose) for several months afterward
  2. Studies on duration of oral anticoagulation for treatment of DVT--3mos may be long enough for first DVT; consider longer anticoagulation if transient risk factors are not present.
  1. 162 pts with 1st DVT or PE w/o known risk factors; all got warfarin (adjusted dose, target INR 2-3) x 3mos then randomized to continue warfarin vs. placebo; study stopped at 10mos b/c of sig. decreased incidence of recurrent DVT in continued-warfarin group (1.3% vs. 27.4%) (NEJM 340:901, 1999--JW)
  2. 267 pts with first-episode of idiopathic DVT all tx'd with 3mos of adjusted-dose (target INR 2-3) warfarin; then randomized to continue x 9mos more vs. d/c; over mean 37mo f/u, no diff. in recurrent venous thromboembolism (16% in each group); nearly all occurred in each group after d/c of anticoagulation.  No diff. in overall mortality rate (NEJM 345:165--JW & abst)
  3. For a second DVT: 227 pts randomized to coumadin x 6mos vs. permanent coumadin; adjusted to INR 2.0-2.8; f/u x 4y showed lower rate of recurrent venous thromboembolism (DVT and/or other), 3% vs. 21%, and nonsig. higher rate of major hemorrhage, 9% vs. 3%. (NEJM 336:393, 1997-JW)
  4. 508 pts with idiopathic DVT or PE randomized to low-intensity warfarin (target INR 1.5-2.0) vs. placebo, long-term; all had received full-dose warfarin anticoagulation (target INR 2.0-3.0) for median 6mos before enrolling.  After avg. 2.1y f/u, incidence of recurrent venous thromboembolism was 5.5% w/warfarin vs. 14.6 w/placebo (sig); no sig. diff. in incidence of major bleeding or death (NEJM 348:1425, 2003--JW)
  5. 738 pts with idiopathic DVT (35% also had PE) who had received at least 3mos of anticoagulation to target INR 2.0-3.0 randomized to continued tx with warfarin at low intensity (target INR 1.5-1.9) vs. conventional intensity (target INR 2.0-3.0).  Over mean 2.4y f/u, low-intensity group had sig. higher rate of recurrent venous thromboembolism (1.9% vs. 0.7%), nonsig. higher overall mortality (1.9% vs. 0.9%), and no sig. diff. in rates of major bleeding (1.1% vs. 0.9%) and any abnormal bleeding (4.9% vs. 3.7%).  No association between carriage of the Factor V Leiden mutation and recurrent venous thromboembolism (NEJM 349:631, 2003--JW)
  6. 1233 pts with prior sympomatic venous thromboembolism (35% had had PE) who had undergone 6mos of warfarin anticoagulation randomized to additional Ximelagatran 24mg BID vs. placebo x 18mos. Incidence of recurrent symptomatic DVT or PE was sig. lower in ximelagatran recipients (3% vs. 13%). Ximelagatran had sig. higher incidence of serum ALT elevations (6% vs. 1%) (NEJM 349:1713, 2003--JW)
  7. In a study in 740 adults with proven or highly probable venous thromboembolism (DVT or PE) randomized to oral anticoagulation x 3mos or 6mos (warfarin, target INR 2.0-3.5), over 12mos, there was no sig. diff. in incidence of (fata or nonfatal thrombotic events) but sig. higher incidence of major hemorrhage in the 6mo group (BMJ 334: 674, 2007--JW)
  8. In a meta-analysis of 15 studies comparing "long-term" (median 6mos) vs. "short-term" (median = 1.75mos) of anticoagulation after venous thromboembolism found that long-term treatment was associated with a sig. reduction in incidence of recurrent venous thromboembolism (NNT = 50) w/o sig. increased risk of bleeding events. (JAMA 294:706, 2005--JW)
  9. In a study in 538 pts with first proximal DVT, all of whom had received anticoagulation x 3mos, randomized to "fixed-duration" anticoagulation (no additional anticoagulation unless DVT was felt to have been "unprovoked" in which case they received another 3mos) vs. "flexible-duration" anticoagulation (all pts underwent LE venous doppler and those with recanalized veins got no additional anticoagulation otherwise got 9mos additional if DVT felt to have been secondary or 21mos if felt to have been unprovoked), over 33mos f/u, incidence of recurrent DVT was sig. lower in "flexible-duration" group (11.9% vs. 17.2%) (Ann. Int. Med. 150:577, 2009-JW)
  1. Ximelagatran for initial treatment of DVT
    1. In a study of 2,489 pts with acute DVT randomized to ximelagatran 36mg PO BID vs. (enoxaparin 1mg/kg SQ BID followed by warfarin to INR 2.0-3.0) x 6mos, there was no sig. diff. in incidence of recurrent VTE, major bleeding, or overall mortality.  Incidence of ALT rising to > 3x upper limit of normal was 9.6% in ximelagatrain group and 2.0% in enoxaparin/warfarin group.  Also, in retrospective analysis of reported adverse events, incidence of severe coronary events was sig. higher in ximalagatrain group (0.81% vs. 0.1%) ("Thrombin Inhibitor in Venous Thromboembolism" ("THRIVE") Trial; JAMA 293:681, 2005--abst)
  1. Using D-dimer level to guide duration of anticoagulation after venous thromboembolism
    1. In a study in 608 adults s/p 3mos of oral anticoagulation for unprovoked, symptomatic venous thrmboembolism, all of whom underwent D-dimer testing 1mo after stopping anticoagulation, with those having elevated levels being randomized to resume anticoagulation vs. stay off it and those with normal D-dimer levels staying off it, over mean 1.4y f/u, among pts with elevated D-dimer, those who restarted anticoagulation had sig. lower incidence of recurrent VTE (3% vs. 15%) and in pts with normal D-dimer levels, incidence of recurrent VTE was 6% (not sig. higher than the continued-anticoagulation group) (NEJM 355:1780, 2006--JW)
  1. Compression stockings after acute phase (may reduce risk of post-DVT sd; see below)
  2. Vena caval filters
    1. 400 pts, avg. age 72yo, with proximal DVT (36% also had symptomatic pulmonary embolus) who were considered at "high risk" for subsequent PE by their docs were randomized to a filter or no filter; all got warfarin x 3mos. Over 2y f/u no sig. diff in rates of symptomatic PE or death; recurrent DVT was higher (21% vs. 12%) in the filter group. (NEJM 338:409, 1998--JW)
      1. In a follow-up report on the same study, at 8y, the filter group had sig. lower incidence of PE (6.2% vs. 15.1%) but sig. higher incidence of recurrent DVT (35.7% vs. 27.5%) ("Prevention du Risque d'Embolie Pulmonaire par Interruption Cave" ("PREPIC") Trial; Circ. 112:416, 2005--JW)
    2. SVC filters after upper-extremity DVT--in an uncontrolled study, they seemed to be effective at preventing PE's in 41 pts with UE DVT who failed anticoagulation or had contraindications to anticoagulation (Radiology 210:53, 1999--JW)

VI. Long-term sequelae:

  1. Recurrence of DVT
    1. Recurrent DVT occurred in 25% at 5y and 30% at 8y in one observational study of 355 pts with 1st-time DVT (Ann. Int. Med. 125:1, 1996)
    2. In one prospective study of 1719 pts with DVT or PE, recurrence rates were 8.3% at 3mos, 10.1% at 6mos, 12.9% at 1y, and 30.4% at 10y. Based on this, the authors actually recc'd 12mos of anticoagulation after a first-episode of DVT (Arch. Int. Med. 160:761, 2000--JW)
    3. Risk factors for recurrence
      1. Age, body mass, paralysis, malignancy, neurosurgery (ibid.)
      2. Residual thrombosis on compression ultrasonography after 3mos of oral anticoagulation was ass'd with sig. higher risk (HR 2.5) for recurrent DVT in a 3-y prospective trial of 313 pts with proximal LE DVT (Ann. Int. Med. 137:955, 2002--JW)
      3. Trauma or surgery before original DVT ass'd with lower recurrence risk
    4. In a prospective study of 610 adults with DVT, with plasma D-Dimer levels measured 3wks after discontinuation of at least 3mos of anticoagulation, over mean 38mo f/u, RR for recurrent venous thromboembolism, after adjustment for potential confounders, was 0.3 (sig.) in pts with D-Dimer levels < 250 ng/mL c/w those with D-Dimer levels of 750ng/mL or greater (JAMA 290:1071, 2003--JW)
  1. "Post-thrombotic syndrome"
    1. A syndrome of various LE sx after DVT, including skin discoloration, edema of varying degrees of severity, chronic pain, and in severe cases, skin ulceration
    2. Often won't manifest for 5-10y after DVT
    3. In an observational study of 355 pts with 1st-time DVT; 245 followed x 5y; 148 followed x 8y (Ann. Int. Med. 125:1, 1996):
      1. Incidence of post-thrombotic syndrome was 28% at 5y and 29% at 8y
      2. Incidence of "severe" post-thrombotic syndrome, e.g. w/ulcer, was 9% at 5y and 9% at 8y
    4. Prevention with graded compression stockings
      1. In a trial of 194 pts randomized to sized-to-fit graded compression stockings vs. no stockings for 2y after DVT, stocking pts had 20% risk of mild-mod post-DVT sd. vs. 47% of controls (Lancet 349:759, 1997-JW)
      2. In a trial of 180 pts with 1st-time DVT, all of whom received oral anticoagulation for 3-6mos, randomized to below-knee graded compression stockings (30-40mm Hg of pressure at ankle) x 2y vs. no stockings; incidence of post-thrombotic syndrome was sig. lower in stocking group (25% vs. 49%); no sig. diff. in 2y incidence of recurrent DVT (Ann. Int. Med. 141:249, 2004--JW)