Post-stroke dysphagia is common, associated with poor outcome and often requires non-oral feeding/fluids. The relationship between route of feeding and outcome, as well as treatment with glyceryl trinitrate (GTN), was studied prospectively. The Efficacy of Nitric Oxide in Stroke (ENOS) trial assessed transdermal GTN (5 mg versus none for 7 days) in 4011 patients with acute stroke and high blood pressure. Feeding route (oral = normal or soft diet; non-oral = nasogastric tube, percutaneous endoscopic gastrostomy tube, parenteral fluids, no fluids) was assessed at baseline and day 7. The primary outcome was the modified Rankin Scale (mRS) measured at day 90. At baseline, 1331 (33.2%) patients had non-oral feeding, were older, had more severe stroke and more were female, than 2680 (66.8%) patients with oral feeding. By day 7, 756 patients had improved from non-oral to oral feeding, and 119 had deteriorated. Non-oral feeding at baseline was associated with more impairment at day 7 (Scandinavian Stroke Scale 29.0 versus 43.7; 2p < 0.001), and worse mRS (4.0 versus 2.7; 2p < 0.001) and death (23.6 versus 6.8%; 2p = 0.014) at day 90. Although GTN did not modify route of feeding overall, randomisation ≤6 h of stroke was associated with a move to more oral feeding at day 7 (odds ratio = 0.61, 95% confidence intervals 0.38, 0.98; 2p = 0.040). As a proxy for dysphagia, non-oral feeding is present in 33% of patients with acute stroke and associated with more impairment, dependency and death. GTN moved feeding route towards oral intake if given very early after stroke. Clinical Trial Registration Clinical Trial Registration-URL: http://www.controlled-trials.com. Unique identifier: ISRCTN99414122.