Reliability of lithium dilution cardiac output in anaesthetized sheep

Cardiac output (CO) measurement with lithium dilution (COLD) has not been fully validated in sheep using precise ultrasonic flow probe technology (COUFP). Sheep generate important cardiovascular research models and the use of COLD has become more popular in experimental settings. Ultrasonic transit-time perivascular flow probes were surgically implanted on the pulmonary artery of 13 sheep. Paired COLD readings were taken at six time points, before and after implantation of a left ventricular assist device (LVAD) and compared with COUFP recorded just after lithium injection. The mean COLD was 5.7 litre min−1 (range 3.8–9.6 litre min−1) and mean COUFP 5.9 litre min−1 (range 4.0–9.2 litre min−1). The bias (standard deviation) was 0.3 (1.0) litre min−1 [5.1 (16.9)%] and limits of agreement (LOA) were −1.7 to 2.3 litre min−1 (−28.8 to 39.0%) with a percentage error (PE) of 34.4%. Data to assess trending [rate (95% confidence intervals)] included a 78 (62–93)% concordance rate in the four-quadrant plot (n=27). In the half moon polar plot (n=19), the mean polar angle was +5°, the radial LOA were −49 to +35° and 68 (47–89)% of data points fell within 22.5° of the mean polar angle. Both tests indicated moderate to poor trending ability. COLD is not precise when evaluated against COUFP in sheep based on the statistical criteria set, but the results are comparable with previously published animal studies.