To evaluate the efficacy of chest roentgenograms and echocardiograms in identifying pulmonary venous pathway obstruction in patients after the modified Fontan procedure, the records of 297 patients who underwent 307 Fontan procedures between 1984 and 1990 were reviewed. Twelve cases of pulmonary venous pathway obstruction documented by autopsy (3 cases) or cardiac catheterization (9 cases with an A wave gradient greater than or equal to 4 mm Hg) were found in 10 patients (mean age 43 +/- 28 months). The mechanisms of obstruction included narrow pulmonary vein ostia in six cases, narrow left atrial outlet in four and atrial baffle obstruction in three. Two causes of obstruction were present in one case. No patient had pulmonary venous congestion on chest roentgenograms. Pathway diameters indexed to 3 square root of body surface area with two-dimensional echocardiography were found to be 6.8 +/- 1.2 mm/3 square root of m2 in the subcostal frontal view, 7.9 +/- 0.6 mm/3 square root of m2 in the subcostal sagittal view and 6.5 +/- 1.7 mm/3 square root of m2 in the apical "four-chamber" view. These values differed significantly from those in 11 age-matched patients undergoing the Fontan procedure without pulmonary venous pathway obstruction documented by catheterization (p = 0.001). With pulsed Doppler ultrasound, there was a relatively narrow range of velocities distal to the obstruction (1.3 to 2.5 m/s). In five of the seven cases with pulsed Doppler measurements, flow was continuous and the Doppler spectral recordings were not phasic. Thus, in patients who have undergone the Fontan procedure and have pulmonary venous pathway obstruction 1) chest roentgenography cannot be used as a screening tool; 2) distal velocities as low as 1.3 m/s occur, usually with nonphasic, continuous forward flow; and 3) pathway diameters indexed to 3 square root of body surface area may be used as an output-independent parameter to cross-check Doppler data.