Reproducibility of double indicator dilution measurements of intrathoracic blood volume compartments, extravascular lung water, and liver function.

STUDY OBJECTIVE: Arterial thermal dye dilution (TDDart) with the COLD system (Munich, Germany) allows measurement of cardiac index (CI), partial blood volumes, lung water, and liver function. The aim of the study was to determine agreement of TDDart measurements with pulmonary artery thermal dilution measurements (TDpa) and to assess the reproducibility of TDDart parameters.

DESIGN: Prospective study.

SETTING: ICU of a university hospital department of cardiac surgery.

PATIENTS: Thirty consecutive patients after coronary artery bypass grafting.

MEASUREMENTS AND RESULTS: Triplicate measurements of TDDart parameters were performed 1, 3, 6, 12, and 24 h postoperatively and coefficients of variation (CVs) were computed. At the 3-h point, additional fivefold TDDart measurements were done and compared with TDpa measurements. The coefficient of correlation for CI from TDDart vs TDpa was 0.96 (p<0.001), and the mean difference was 0.16 L/min/m2 (2.4%). The CVs of the TDDart and TDpa CI measurement were 7.2% and 5.9%; the CVs of other TDDart parameters were 4.6% (cardiac function index), 8.3% (global end-diastolic volume), 7.0% (intrathoracic blood volume), 7.6% (total blood volume), 7.4% (right ventricular end-diastolic volume), 7.4% (right heart end-diastolic volume), 11.3% (left heart end-diastolic volume [LHEDV]), 12.0% (right to left heart volume proportion [R/LHV]), 8.8% (pulmonary blood volume), 10.8% (extravascular lung water), 16.4% (plasma disappearance rate of dye), and 19.8% (dye clearance). The CV did not depend on Glasgow coma scale or on body temperature.

CONCLUSION: The CVs of LHEDV and R/LHV are influenced by asynchronous TDDart and TDpa variation. The CVs of plasma disappearance and dye clearance are increased as the half-life of the dye is longer than the measurement sequence. All other parameters derived from TDDart and TDpa show a clinically sufficient reproducibility.