Objectives To assess variation in feeding practice at hospital discharge after the Norwood procedure, factors associated with tube feeding, and associations among site, feeding mode, and growth before stage II. Study design From May 2005 to July 2008, 555 subjects from 15 centers were enrolled in the Pediatric Heart Network Single Ventricle Reconstruction Trial; 432 survivors with feeding data at hospital discharge after the Norwood procedure were analyzed. Results Demographic and clinical variables were compared among 4 feeding modes: oral only (n = 140), oral/tube (n = 195), nasogastric tube (N-tube) only (n = 40), and gastrostomy tube (G-tube) only (n = 57). There was significant variation in feeding mode among sites (oral only 0%-81% and G-tube only 0%-56%, P < .01). After adjusting for site, multivariable modeling showed G-tube feeding at discharge was associated with longer hospitalization, and N-tube feeding was associated with greater number of discharge medications ( R2 = 0.65, P < .01). After adjusting for site, mean pre–stage II weight-for-age z -score was significantly higher in the oral-only group (−1.4) vs the N-tube–only (−2.2) and G-tube–only (−2.1) groups ( P = .04 and .02, respectively). Conclusions Feeding mode at hospital discharge after the Norwood procedure varied among sites. Prolonged hospitalization and greater number of medications at the time of discharge were associated with tube feeding. Infants exclusively fed orally had a higher weight-for-age z score pre–stage II than those fed exclusively by tube. Exploring strategies to prevent morbidities and promote oral feeding in this highest risk population is warranted.