Crystalloid versus colloid for fluid resuscitation of hypovolemic patients.
Academic Article
Overview
abstract
The choice of the initial asanguinous fluid--either crystalloid or colloid--used for the resuscitation of the hypovolemic patient remains controversial. Colloid supporters argue for the careful preservation of the plasma colloid osmotic pressure (PCOP) to protect the lung from pulmonary edema. A careful analysis of the Starling microvascular forces operative at the pulmonary capillary makes such an effect unlikely. In fact, the lung is relatively immune to hemodilution and any decrease in PCOP is roughly one fourth as important as increases in hydrostatic pressure in causing increased fluid exchange. A critical review of the experimental and clinical studies comparing crystalloid versus colloid resuscitation essentially shows no physiologic difference in the two solutions. Using the thermal-green dye technique of extra-vascular lung water (EVLW) measurement in twenty crystalloid resuscitated trauma (n = 10) and burn (n = 10) patients, we have specifically evaluated the pulmonary effects of profound depression of PCOP and a negative PCOP - PAWP gradient (a shorthand form of the Starling equation argued to predict the presence of pulmonary edema if + 4 mm Hg or less). Average resuscitative fluid volumes during the first two hospital days were: 31.8 litres of crystalloid and no colloid for each burn patient; and 18.5 liters of crystalloid, 21 units of blood and 1 liter of colloid (as fresh frozen plasma) for each trauma patient. EVLW remained in the normal range of 7.0 +/- 1.0 ml/kg during the first five hospital days for all patients despite profound decrease in PCOP (less than 15 mm Hg) and a low or negative PCOP - PAWP gradient. Crystalloid resuscitation clearly is not harmful to the lung and it is equally as effective as colloid resuscitation. Crystalloid is markedly less expensive than colloid and, given the greater cost of colloid without evident benefit, one wonders how their further use can be justified.
