Effects of one-lung ventilation on thermodilution-derived assessment of cardiac output.

BACKGROUND Cardiac output (CO) monitoring can be useful in high-risk patients during one-lung ventilation (OLV), but it is unclear whether thermodilution-derived CO monitoring is valid during OLV. Therefore, we compared pulmonary artery (CO(PATD)) and transcardiopulmonary thermodilution (CO(TPTD)) with an experimental reference in a porcine model. METHODS CO(PATD) and CO(TPTD) were measured in 23 pigs during double-lung ventilation (DLV) and 15 min after the onset of OLV, during conditions of normovolaemia and after haemorrhage. An ultrasonic flow probe placed around the pulmonary artery (CO(PAFP)) was used for reference. RESULTS The range of CO in these experiments was 1.5-3 litre min(-1). Normovolaemia: during DLV and conditions of normovolaemia, the mean (95% limits of agreement) bias for CO(PATD) compared with CO(PAFP) was -0.05 (-0.92 and 0.83) litre min(-1), and 0.58 (-0.40 and 1.55) litre min(-1) for CO(TPTD). During OLV, the bias for CO(PATD) remained unchanged at 0.08 (-0.51 and 0.66) litre min(-1), P=0.15, and the bias for CO(TPTD) increased significantly to 0.85 (0.05 and 1.64) litre min(-1), P=0.047. Hypovolaemia: during DLV, the bias for CO(PATD) compared with CO(PAFP) was 0.22 (-0.20 and 0.66) litre min(-1) and for CO(TPTD) was 0.60 (0.12 and 1.10) litre min(-1). There was no significant change of bias during OLV for CO(PATD) [0.30 (-0.10 and 0.70) (litre min(-1)), P=0.25] or bias CO(TPTD) [0.72 (0.21 and 1.22) (litre min(-1)), P=0.14]. Trending ability during OLV, quantified by the mean of angles θ, showed good values for both CO(PATD) (θ=11.2°) and CO(TPTD) (θ=1.3°). CONCLUSIONS CO(TPTD) is, to some extent, affected by OLV, whereas CO(PATD) is unchanged. Nonetheless, both methods provide an acceptable estimation of CO and particularly of relative changes of CO during OLV.