ObjectiveTo systematically investigate the reproducibility and reliability of a newly developed, less invasive approach of estimating blood volume (BV), using indocyanine green (ICG) measured with pulse-spectrophotometry.DesignProspective, clinical study.SettingSurgical unit at a university hospital.PatientsTwenty-two patients undergoing general anesthesia for elective surgery and seven healthy volunteers.InterventionsCatheters were inserted into the forearm veins of healthy volunteers for the administration of ICG and blood sampling for the measurement of hemoglobin concentration.Measurements and Main ResultsThe distribution volumes of ICG in seven healthy volunteers were estimated repetitively following three or four consecutive intravenous administrations at 30-min intervals. A low intrasubject coefficient of variation of 3.94 +/- 2.03 (SEM) % and a reasonable intersubject coefficient of variation of 13.3 +/- 5.52% (in mL/kg) for the BV measurements were obtained. In addition, ICG was administered to 22 patients, first under general anesthesia by a bolus, and then by a bolus with a constant-rate infusion. The ICG blood concentration was noninvasively measured with pulse-spectrophotometry. The blood concentration time courses following both bolus and constant-rate infusion were well fitted by the one-compartment model, indicating that the distribution equilibrium of ICG is instantaneous. The distribution volumes estimated following bolus injection correlate closely with the distribution volume estimated based on constant-rate infusion administration (r2= .90).ConclusionsThe BV estimation with a bolus injection of ICG and pulse-spectrophotometry is reliable, as reflected by the reproducible BVs estimated in the same subject. The integrated pulse-spectrophotometry monitoring system offers a less invasive and useful tool for bedside estimation of BV. (Crit Care Med 1998; 26:1446-1451)