Gastrointestinal permeability was evaluated in 90 fasting healthy subjects and 60 patients with food allergy by oral administration to both groups of 5 g of mannitol, a marker of absorption of small molecules, and 5 g of lactulose, a marker of abnormal absorption of large molecules, and subsequent measurement of urinary excretion of mannitol and lactulose. In healthy subjects, mean 5-hour urinary excretion of mannitol was 14.11% and of lactulose 0.26%. In the fasting state, the 60 patients with food allergy exhibited a mean urinary recovery of mannitol of 13.22%, not significantly different from that in healthy subjects. Mean recovery of lactulose in the patients with food allergy was 0.55%, significantly greater than in the healthy patients. After ingestion of food allergens by the patients, mean mannitol recovery fell to 11.57% and mean recovery of lactulose rose to 1.04%, both values being significantly different from those obtained in the fasting patients. On challenging the patients after they had taken sodium cromoglycate, mean mannitol and lactulose recoveries (13.53% and 0.62%, respectively) were not significantly different from those in fasting patients but were significantly different from those obtained on challenging patients unprotected by sodium cromoglycate. Evaluation of intestinal permeability in this way provides an objective means of diagnosing food allergy and assessing the effectiveness of anti-allergic agents such as sodium cromoglycate.