The minimally invasive CardioQP oesophageal Doppler probe estimates cardiac output by measuring blood flow velocity in the descending aorta. Individual variables to enter are patient's age, weight and height. We measured cardiac output simultaneously with CardioQP and pulmonary artery catheter thermodilution techniques during heart catheterisation in 40 paediatric patients with congenital heart defects. Median [range] age was 8.2 years [0.5-16.7 years], cardiac output values measured by thermodilution and CardioQP were 3.6 l.min(-1) [1.2-7.1 l.min(-1)] and 3.0 l.min(-1) [0.7-6.7 l.min(-1)], respectively. These values showed only moderate correlation (r = 0.809; p < 0.0001). Bias and precision were 0.66 l.min(-1) and 1.79 l.min(-1) (95% limits of agreement: -1.13 to +2.45 l.min(-1)). Based on our preliminary experience, cardiac output values measured by CardioQP in children do not reliably represent cardiac output values compared with the thermodilution technique. We suggest measurement of individual aortic diameter to improve performance of the CardioQP.