The maximum cross-sectional area of the central pulmonary arteries indexed to the body surface area (pulmonary artery index [PAI]) was measured preoperatively from angiograms in 173 patients evaluated for a Fontan-type operation between 1981 and June 1989. Of these, 34 patients underwent another palliative procedure, 8 primarily on the basis of small pulmonary arteries (PAI, 106 to 167 mm2./m2) . The mean PAI of this group was significantly smaller than the mean PAI of the remaining 139 patients who underwent a Fontan operation (136 +/- 20 versus 310 +/- 113 MM2/M2) (p < 0.001). The patients who underwent a Fontan operation were evaluated according to three overlapping end points: (1) hospital death or takedown of repair (12.2%), (2) early failure (cumulative death or takedown of repair within 6 months of operation) (16.5%), and (3) early failure or persistent effusions (33.8%). With regard to these end points, no significant difference in pulmonary artery size could be found between patients having a favorable or unfavorable outcome. However, among a low-risk subset of 30 patients with tricuspid atresia, those with ''early failure or persistent effusions'' had significantly smaller pulmonary arteries than those with a good outcome (PAI, 185 +/-47 versus 276 +/- 83 MM2/M2 ) (p < 0.01). The postoperative transpulmonary gradient of the 8 patients with the smallest pulmonary arteries who underwent a Fontan operation (all PAIs < 170 MM2/m2) was significantly greater than that of the rest of the study group (9.88 +/- 2.3 versus 8.13 +/- 2.3 mm Hg) (p < 0.04). Small pulmonary arteries, in otherwise good-risk patients, are associated with increased risk of ''early failure or persistent effusions'' after the Fontan operation. The lowest PAI compatible with success remains unknown. To obtain a more comprehensive assessment of a successful outcome, we suggest that evaluation of patients after a Fontan operation include end points other than hospital death.
