OBJECTIVE: To determine whether the proximal isovelocity surface area (PISA) method could be applied to estimate the magnitude of ventricular septal defect (VSD) shunt flow. DESIGN: Prospective analysis of clinical, echocardiographic, and angiographic data. SETTING: University hospital. PATIENTS: 14 children with VSD. METHODS: Colour Doppler images of VSD shunt flow were obtained in parasternal long axis view, four chamber view or both, adjusted to provide the best imaging of flow. The VSD shunt flow rate and shunt volume were calculated as follows: shunt flow rate (SFR) = 2 pi r2 V/BSA in ml/s/m2; shunt volume = SFR x shunt duration time. The shunt volume, shunt fraction, and pulmonary to systemic flow ratio (Qp:Qs) were confirmed by cardiac catheterisation. RESULTS: There was a correlation between shunt variables determined by PISA and those by catheterisation, including shunt volume (r = 0.78, P = 0.001) and shunt fraction (r = 0.74, P = 0.003). Qp:Qs was also significantly correlated with SFR (r = 0.79, P = 0.0007). The SFR was significantly different between the four patients with Qp:Qs < 2.0 (mean (SD) 54 (33) ml/s/m2) and the 10 patients with Qp:Qs > 2.0 (186 (69) ml/s/m2) (P = 0.004). CONCLUSIONS: These data suggest that the PISA method is a reliable non-invasive investigation for the quantitative assessment of VSD shunt flow and provides important information for decisions regarding surgical repair.