Measurement of cardiac output: a comparison between transpulmonary thermodilution and uncalibrated pulse contour analysis

Background: Recently, continuous monitoring of cardiac output (CO) basedon pulse contour analysis (Vigileo^®) has been introducedinto practice. In this clinical study, we evaluated the accuracyof this system by comparing it with the transpulmonary thermodilutiontechnique (TPID) in septic patients. Methods: We studied 24 mechanically ventilated patients with septic shock(16 male, 8 female, age 26–77 yr) receiving treatmentwith norepinephrine who for clinical indication underwent haemodynamicmonitoring by the transpulmonary thermodilution technique usinga PiCCO^®plus system (Pulsion Medical Systems, Munich, Germany).In parallel, arterial pulse contour was applied using the femoralarterial pressure curve (FloTrac^® pressure sensor, Vigileo^®monitor, Edwards Lifesciences, Irvine, USA). After baselinemeasurement, mean arterial pressure was elevated by increasingnorepinephrine dosage, and CO was measured again before meanarterial pressure was reduced back to baseline levels. Fluidstatus and ventilator settings remained unchanged throughout.At each time point, CO by transpulmonary thermodilution wascalculated from three central venous bolus injections of 15ml of saline (<8°C). Linear regression and the Bland–Altmanmethod were used for statistical analysis. Results: Overall, CO was 6.7 (SD 1.8) (3.2–10.1) litre min^–1for CO(TPID) and 6.2 (2.4) (3.0–17.6) litre min^–1for CO(Vigileo^®). Linear regression revealed: CO(Vigileo^®)= 1.54 + 0.72 x CO(TPID) litre min^–1, r² = 0.26 (P <0.0001). Mean bias between techniques CO(TPID)–CO(Vigileo^®)]was 0.5 litre min^–1 (SD 2.3 litre min^–1). Correlationcoefficients at the three time points were not significantlydifferent from each other. Conclusions: Pulse contour analysis-derived CO (Vigileo^® system) underestimatesCO(TPID) and is not as reliable as transpulmonary thermodilutionin septic patients.