Cjb. Wallin et al., QUANTITATIVE ESTIMATION OF ERRORS IN THE INDICATOR DILUTION MEASUREMENT OF EXTRAVASCULAR LUNG WATER, Intensive care medicine, 23(4), 1997, pp. 469-475
Objective: To assess the accuracy of the diffusible indicators heavy w
ater and thermal indicator in the measurement of extravascular lung wa
ter (EVLW). Design: Cardiac output (GO), mean transit time and EVLW fo
r the two diffusible indicators were measured. CO for indocyanine gree
n, gravimetric EVLW and the calculated mean transit time for a diffusi
ble indicator were used as independent reference variables. CO, mean t
ransit time and EVLW for the two diffusible indicators were compared t
o the reference variables and the percentage error for each measured v
ariable was calculated for each bolus injection. Setting: 6 sheep with
healthy lungs and 6 with pulmonary oedema in a research laboratory. I
nterventions: CO was altered with positive end-expiratory pressure and
dobutamine. Measurements: All indicators were given together in a bol
us through a central venous line. Indicators were detected simultaneou
sly in the aorta, and CO, mean transit time and EVLW were measured. EV
LW was measured gravimetrically (EVLWgrav) postmortem. Results: In the
combination of a low CO and a large distribution volume, heavy water
and thermal indicator produced a large number of slow wash-out curves.
These curves were abolished from further analysis. The mean errors in
CO and mean transit time for heavy water were close to zero and indep
endent of the distribution volume; the product EVLW was close to EVLWg
rav. The mean error in thermodilution CO measured in the aorta was clo
se to zero but dependent on the distribution volume. The mean error in
mean transit time for the thermal indicator was 36 % and dependent on
the distribution volume. Their product EVLW overestimated EVLWgrav by
70 %. Conclusions: The results obtained for heavy water confirmed the
theoretical basis of the indicator dilution method. The mean transit
time for the thermal indicator was not proportionate to its distributi
on volume. The magnitude of this error prevents the calculation of an
anatomically defined EVLW using a catheter-mounted thermistor in the a
orta.