Using laser Doppler flowmetry, we measured gastric and duodenal mucosal blood flow in 70 patients who had taken non-steroidal anti-inflammatory drugs (NSAIDs) for longer than 4 weeks, and studied the correlation with demographic factors, ulceration, and Helicobacter pylori. Blood flow was also measured in 17 other subjects not taking any drugs. Measurements were taken from healthy-looking mucosa in the gastric antrum and the first part of the duodenum. Both gastric and duodenal blood flow values were significantly lower in patients taking NSAID than in those who did not. In the NSAID group, the median duodenal mucosal blood flow was 150 perfusion units in smokers (n = 29) compared with 175 in non-smokers (P = 0.024), 123 units in patients with duodenal ulcers (n = 12) compared with 160 in those without duodenal ulcers (P = 0.020), 135 units in patients with H. pylori (n = 30) compared with 168 in patients without H. pylori (P = 0.033), and 118 in smokers infected with H. pylori compared with 175 units in non-smokers not infected with H. pylori (F = 13.4, P = 0.0005). There was no correlation with age. Gastric blood flow was not significantly influenced by any of the above variables. These results suggest that chronic NSAID intake is associated with reduced blood flow in both the stomach and duodenum. However, amongst NSAID patients, duodenal, but not gastric, mucosal blood flow is reduced in smokers, and in those with duodenal ulcers and H. pylori.(ABSTRACT TRUNCATED AT 250 WORDS)