Effects of crystalloid solutions on circulating lactate concentrations: Part 1. Implications for the proper handling of blood specimens obtained from critically ill patients.

Abstract

a) To test the hypothesis that circulating lactate concentrations are the same in simultaneously collected arterial and central venous blood specimens; b) to test the hypothesis that even small amounts of crystalloid solutions, which are inadequately "cleared" from these indwelling arterial and venous catheters, can lead to clinically important and misleading changes in the measured lactate values. A prospective, multiexperiment study. A critical care research laboratory and a 20-bed intensive care unit (ICU). Three hundred fifty-five patients. Blood samples were collected. Experiment 1: Simultaneously collected arterial and central venous blood specimens were obtained on 148 occasions from 48 medical ICU patients receiving no lactated Ringer's solution (RL). Arterial and central venous lactate values were nearly identical in these patients. The correlation between the arterial and central venous lactate concentrations was excellent (r2 = .85; p < .0001) and the agreement between the arterial and central venous lactate concentrations was also excellent (bias and precision = 0.04 mmol/L and +/- 0.38 mmol/L, respectively). Experiment 2: Arterial and mixed venous blood samples were obtained from 100 percutaneous transluminal coronary angioplasty (PTCA) and 75 cardiac surgical patients immediately before the performance of these cardiac procedures. We found the central venous lactate concentrations to be higher than arterial lactate values in the cardiac surgical group, and there was a very poor correlation (r2 = .07) between arterial and central venous lactate values in the cardiac surgical group. The correlation between central venous and arterial lactate concentrations in the PTCA patients was excellent (r2 = .84) and similar to the findings of experiment 1. Since the cardiac surgical patients received RL and the PTCA patients received no RL, we speculated that the intravenous infusion of RL in the cardiac surgical group accounted for these discordant findings. To test this speculation, we performed experiments 3 and 4. Experiment 3: In a large bench study, blood specimens were divided into multiple 1-mL aliquot portions, to which 0.01, 0.05, 0.10, 0.50, or 1.0 mL of various crystalloid solutions, containing or not containing RL, were added. In a volume-dependent and linear manner, solutions containing RL increased the circulating lactate concentration from 10% to > 400% of the baseline lactate value. In a volume-dependent and linear fashion, the non-RL crystalloid solutions decreased the lactate concentration by 0 to 66% of the baseline nondiluted lactate concentration. Experiment 4: In 30 different cardiac surgical patients, we simultaneously obtained central venous and arterial blood specimens. Patients this time received no RL, and catheter lines were adequately cleared (removal > 5 mL) of crystalloid solutions. We found a correlation (r2 = .82; p < .0001) that was virtually identical to the findings of experiment 1 and to the findings in the PTCA group of experiment 2. a) Arterial and central venous lactate concentrations are similar in hemodynamically stable critically ill patients, b) Even small amounts of RL-containing solutions in catheters used for blood sampling may cause false increases in the circulating lactate concentration. c) Even small amounts of non-RL crystalloid solutions in catheters used for blood sampling may falsely decrease circulating lactate values. d) When blood specimens are drawn from indwelling catheters, all crystalloid solutions must be cleared from the line.