Deep vein thrombosis: ultrasound is better than impedance plethysmography.

Clinical bottom line (level 1a)

  1. Ultrasound scanning is better than IPG at diagnosing and excluding deep vein thrombosis in patients with symptoms.
  2. Serial IPG or ultrasound scan can safely exclude DVT.
  3. Neither is very good at diagnosing or excluding DVT in asymptomatic patients.
  4. Neither is very good at excluding calf DVT.
Kearon et al: Annals of Internal Medicine 1998; 128 (8): 663-677
Expires June 2003

The study

Systematic review of all of
  • Patients: suspected deep vein thrombosis
  • Intervention: impedance plethysmography and ultrasound scan
  • Outcome: diagnosis


    • Articles found in ? using MEDLINE, 1966 to January 1997 (search terms: thrombosis, thrombophlebitism plethysmography, impedance ultrasonography, randomised controlled trials and cohort studies ) and bibliographic searches

    Selection criteria: as above
    Appraisal criteria: appraised using set criteria. 10% of articles were reappraised by second reviewer (94% correlation). Evidence and recommendations were graded.
    Articles excluded if:

    it is unclear how many studies were included.

    The evidence


    diagnostic test symptomatic DVT/ asymptomatic/ abnormal scan no symptomatic DVT/asymptomatic/abnormal scan LR+
    (95% CI)    		 post-test probability LR-
    (95% CI)    		 post-test probability
    symptomatic- impedance plethysmography 248 62 11
    (8.7 to 14)     		80% 0.20
    (0.16 to 0.25)     		7%
    symptomatic- ultrasound scan 926 66 23
    (18 to 29)     		93% 0.08
    (0.06 to 0.10)     		5%
    asymptomatic- IPG 52 46 3.8
    (2.6 to 5.6)     		53% 0.88
    (0.84 to 0.92)     		21%
    asymptomatic- ultrasound scan 229 63 9.8
    (7.6 to 13)     		78% 0.60
    (0.56 to 0.65)     		9.6%
    total . .

    • Pre-test probabilities:
      • symptomatic- impedance plethysmography: 26%
      • symptomatic- ultrasound scan: 39%
      • asymptomatic- IPG: 23%
      • asymptomatic- ultrasound scan: 27%
    • serial IPG:
      • abnormal scan (95% CI): 17% (14% to 20%)
      • conversion during serial testing: 6.1% (1.0% to 11%)
      • missed venous thromboembolism: 1.5% (0.8% to 2.2%)
    • serial ultrasound scan:
      • abnormal scan: 31% (17% to 44%)
      • conversion during serial testing: 1.1% (0.7% to 1.5%)
      • missed venous thromboembolism: 2.0% (0.0% to 4.9%)
    • There was significant heterogeneity between trials with regard to serial IPG.

    Comments

    1. Likelihood ratios were calculated from raw data presented- not weighted, but unlikely to be very different.
    2. Significant variation between trials was found, eg. two recent studies on IPG have suggested it is less good than originally thought.

    Citation

    1. Kearon C, Julian JA, Newman TE, et al: Non-invasive diagnosis of deep venous thrombosis. Annals of Internal Medicine 1998; 128 (8): 663-677
    Contributor: Chris Ball and Clare Wotton, June 2000
    Reviewer: Alex Gallus

    Clinical Question.
    Patient suspected DVT
    Intervention or Exposure impedance plethysmography and ultrasound scan
    Outcome diagnosis