Heart failure: pulmonary blood flow, vascular pedicle width and pulmonary oedema can helped distinguish it from ARDS and overhydratio.n

Clinical bottom line (level 2b)

  1. Half of patients with pulmonary oedema in intensive care units had heart failure.
  2. Patients with inverted pulmonary blood flow were more likely to have heart failure (LR+14) , and patients with patchy pulmonary oedema were less likely to have heart failure (LR+0.0) .
  3. Patients were more likely to have ARDS if they have patchy pulmonary oedema or a narrow vascular pedicle width (LR+11) .
  4. Patients were more likely to be overhydrated if they have a wide vascular pedicle width (LR+6.5) or central pulmonary oedema (LR+5.7) , and less likely to be overhydrated if they had normal pulmonary blood flow or a narrow vascular pedicle width.
Milne et al: American Journal of Roentgenology 1985; 144: 879-894
Expires March 2003

The study

Setting: two intensive care units, Italy and USA

118 patients (aged ARDS, ?% male) with radiographic evidence of pulmonary oedema

Excluded if
  • no clinical or physiological documentation of the cause of heart failure



    • Independent blinded reference standard, applied in all patients from a consecutive inappropriate spectrum.
    Reference standard:
    • physiological evaluation, and histology in some cases
    Diagnostic test: chest X-ray
    • pulmonary blood flow distribution: normal (lower lung fields predominate); balanced (equal in lower and upper lung fields); inverted (upper lung fields predominate)</item><item> distribution of pulmonary oedema: even (homogeneous from chest wall to heart and obeys gravity); central (perihilar only); patchy (spares many areas and often displays air bronchograms)
    • vascular pedicle width (just above aortic knob): narrow (&lt; 43 mm); normal (43-53 mm); wide (&gt; 53 mm)

    The evidence

    pre-test probability of heart failure: 51%, (95% CI: 42% to 60%)
    pre-test probability of ARDS: 24%, (95% CI: 16% to 31%)
    pre-test probability of overhydration: 25%, (95% CI: 17% to 33%)

    diagnostic test heart failure no heart failure LR
    (95% CI)    		 post-test probability
    inverted pulmonary blood flow 29 2 14
    (3.4 to 55)     		94%
    balanced pulmonary blood flow 23 38 0.58
    (0.40 to 0.84)     		38%
    normal pulmonary blood flow 9 18 0.48
    (0.23 to 0.97)     		33%
    total 61 58


    diagnostic test heart failure no heart failure LR
    (95% CI)    		 post-test probability
    even pulmonary oedema 56 16 3.3
    (2.2 to 5.1)     		78%
    central pulmonary oedema 5 30 0.16
    (0.066 to 0.38)     		14%
    patchy pulmonary oedema 0 12 0.0
    (0.0 to 0.24)     		0%
    total 61 58


    diagnostic test heart failure no heart failure LR
    (95% CI)    		 post-test probability
    wide vascular pedicle width (>53 mm) 35 33 1.0
    (0.74 to 1.4)     		51%
    normal vascular pedicle width (43-53 mm) 23 15 1.5
    (0.85 to 2.5)     		61%
    narrow vascular pedicle width (<43 mm) 3 10 0.29
    (0.083 to 0.98)     		23%
    total 61 58


    diagnostic test no ARDS LR
    (95% CI)    		 post-test probability
    inverted pulmonary blood flow 2 29 0.22
    (0.057 to 0.88)     		6%
    balanced pulmonary blood flow 14 47 0.97
    (0.64 to 1.5)     		23%
    normal pulmonary blood flow 12 15 2.6
    (1.4 to 4.9)     		44%
    total 28 91


    diagnostic test ARDS no ARDS LR
    (95% CI)    		 post-test probability
    even pulmonary oedema 10 62 0.52
    (0.31 to 0.88)     		14%
    central pulmonary oedema 6 29 0.67
    (0.31 to 1.5)     		17%
    patchy pulmonary oedema 12 0 -
    (13 to infinity)     		100%
    total 28 91


    diagnostic test ARDS no ARDS LR
    (95% CI)    		 post-test probability
    wide vascular pedicle width 8 60 0.43
    (0.24 to 0.79)     		12%
    normal vascular pedicle width 10 28 1.2
    (0.65 to 2.1)     		26%
    narrow vascular pedicle width 10 3 11
    (3.2 to 37)     		77%
    total 28 91


    diagnostic test overhydration no overhydration LR
    (95% CI)    		 post-test probability
    inverted pulmonary blood flow 6 31 0.57
    (0.27 to 1.3)     		16%
    balanced pulmonary blood flow 24 37 1.9
    (1.4 to 2.6)     		39%
    normal pulmonary blood flow 0 21 0.0
    (0.0 to 0.40)     		0%
    total 30 89


    diagnostic test overhydration no overhydration LR
    (95% CI)    		 post-test probability
    even pulmonary oedema 6 66 0.27
    (0.13 to 0.56)     		8%
    central pulmonary oedema 24 11 6.5
    (3.6 to 12)     		69%
    patchy pulmonary oedema 0 12 0.0
    (0.0 to 0.70)     		0%
    total 30 89


    diagnostic test overhydration no overhydration LR
    (95% CI)    		 post-test probability
    wide vascular pedicle width 25 13 5.7
    (3.4 to 9.7)     		66%
    normal vascular pedicle width 5 33 0.45
    (0.19 to 1.1)     		13%
    narrow vascular pedicle width 0 43 0.0
    (0.0 to 0.20)     		0%
    total 30 89

    Comments

    1. Lung volume, heart shape, widening of fissures and hilar abnormalities were not found to be useful in distinguishing the three types of oedema.

    Citation

    1. Milne EN, Pistolesi M, Miniati M, et al: The radiologic distinction of cardiogenic and non cardiogenic edema. American Journal of Roentgenology 1985; 144: 879-894
    Contributor: Chris Ball and Clare Wotton, March 2001
    Reviewer:

    Clinical Question.
    Patient intensive care with pulmonary oedema
    Intervention or Exposure chest X-ray
    Outcome ARDS, heart failure, overhydration