In the measurement of flow by dye-dilution techniques, the optical density of the dyeblood mixture must stabilize in the interval between dilution and detection if an accurate measurement of flow is to be obtained. For many years indocyanine green (ICG), the most widely used dye in hemodynamic studies, was thought to sbabilize within 1.5 to 2 sec after dilution (1). However, a recent study ( 2 ) failed to reproduce these results. Another study ( 3 ) concluded that slow optical stabilization of ICG results in marked overestimation of flow if appearance times of the dilution curves are under 6 sec, a common occurrence in many applications. Others (4) have also found errors suggestive of slow stabilization. Since multiple potential sources of error complicate these studies, and since the good agreement reported between direct flow measurements and values determined by ICG (5-7) involved long stabilization times, we studied the optical stabilization of ICG direutly in order to define the circumstances in which this could cause errors in flow measurement.Materials and Methods. Concentrated solutions of ICG2 at 1 mg/ml were prepared in media varying in protein and ionic composition, shown in detail in Fig. 2, by adding various solutions to aliquots of a 2.5 mg/ml solution of ICG in (the aqueous diluent supplied. Pllasma from fasted dogs, anticoagulated with ACD or heparin, was used with or without sodium phosphate buffers. 1Thk study was supported in lyart by Public Health Service Grant l3L 08068 and gmnts from the Gradiuate Schiod of the University of Minnesota and the Minnesota Heart Association. 2Oardiogren dye, lots 222 and 658, Hynson, Wevtcott land Dunning, Inc., Bldtimo1re, MD.Changes in optical densilty (OD) of ICG after dilution in plasma were continuously measured on a split-beam, dual-wavelength recording spectr~photometer.~ Mixing-dilution was accomplished with la plunger a~sernbly,~ mounted on top of l a 1 cm cuvette, consisting of a cup on a spring-loaded shaft which on rapid depression into the cuvetrte, created turbulence, mixing the ICG in the cup with the plasma in the cuvette before re-emerging. Nonspecific OD changes, for example due to bubble formation during mixingd ilultion, necessitated use of the dualwavelength mode (8). In this mode light beams at two wavelengths are alternately passed through the cuvette, and the difference in OD of the cuvetlte contents at these wavelengths is recorded. Wavelengths used were 805 nm, which is in the region used for detection of ICG in hemdynamitc applications, and 890 nm, where ICG has no significant absorbsance (9-1 1 ) . The degree of compensation for nonspecific OD changes thus achieved was (assessed by repeating the plunger depressions after the OD had stabilized following diilution. With this method, the duration of the mixing artifact was always less than 1.6 sec, so that OD cbanges due only to ICG could be measured 2 sec after dilution.Results are reported in percentage of stabilized OD because ist was found that a variable amount of dye was emptied fr...