Summary and conclusionsGastric emptying after a test meal was studied in 17 normal volunteers-10 habitual smokers and seven non-smokers. The solid component of the test meal was labelled with technetium and the liquid component with indium. After one meal the habitual smokers smoked two cigarettes. Emptying curves were produced for both technetium and indium, and the differences between curves for meals with and without cigarettes were analysed.Cigarette smoking accelerated the rate at which the liquid component of a meal left the stomach. This may be important in the pathogenesis of duodenal ulcer and the delay in healing caused by cigarette smoking. IntroductionCigarette smoking may be important in the pathogenesis of duodenal ulcer and appears to delay healing.' The intermediary mechanism, however, is not clear. The concept that the rate of gastric emptying tends to be more rapid than normal in patients with duodenal ulcer23 is controversial4 5; if true, this phenomenon may be either a cause or effect of the disease. We think it is probably causative, and a rapid rate of gastric emptying may be an important factor in the pathogenesis of duodenal ulcer.6 Because of this we investigated the effect of cigarette smoking on gastric emptying in normal subjects. We thought that cigarette smoking and duodenal ulcer might be linked, in which case smoking a cigarette would probably accelerate gastric emptying. By using radioactive tracer techniques that had recently become available 5-7 we were able to measure simultaneously the rates at which the solid and liquid components of a meal left the stomach.

SUMMARY We studied the rates at which solid and liquid leave the stomach after meals of wholemeal and white bread by using a double isotope technique. There was no difference in the rates at which the solid phases of the gastric contents left the stomach but liquid left the stomach significantly more rapidly with white bread than with wholemeal bread. Furthermore, the amount of liquid leaving the stomach unaccompanied and therefore unbuffered by solid was significantly greater after white bread than wholemeal bread. These findings may be of significance in the pathogenesis of duodenal ulcer and they provide a rational basis for a possible form of dietary treatment.Carbohydrate refinement has been proposed to be of importance in the aetiology of duodenal ulcer (Cleave, 1962(Cleave, , 1974, but the suggestion that this might be due to loss of food buffer in the refining process has not been supported by experimental work (Tovey, 1974).It has been suggested (George, 1968;Griffith et al., 1968)-although disputed (Cobb et al., 1971;Heading et al., 1976)-that patients with duodenal ulcer tend to have a more rapid rate of gastric emptying than normal but even if this phenomenon is real it is not known whether it is a cause or effect of the disease. We thought that the effect of carbohydrate refinement on gastric emptying in the normal subject should be investigated, as we thought it conceivable that herein might lie a possible link between carbohydrate refinement and duodenal ulcer. We expected that if this were so then gastric emptying would be more rapid after refined carbohydrate than after its unrefined counterpart.Previous studies of the rates at which refined and unrefined carbohydrate leave the stomachs of normal subjects have produced conflicting results (McCance et al., 1953;Wilson et al., 1929) and, furthermore, they involved the measurement of the rate at which barium sulphate given with food left the stomach. By using radioactive tracer techniques which have recently become available (Griffith et al., 1966;Harvey et al., 1970) Received for publication 16 February 1977 measure simultaneously the rates at which both the solid and the liquid components of a meal are emptied from the stomach. MethodWe studied II normal subjects, nine male and two female, aged between 20 and 30 years. All gave consent having been informed in detail of the nature of the investigation. The two females had been sterilised. Each ate meals of wholemeal and white bread in randomised sequence at intervals of one week. The meals were taken after an overnight fast.The bread was prepared in the laboratory using a standard recipe of 120 g flour, 2 g sugar, 1 g salt, 1 g dried yeast, 1-5 ml corn oil, 25 ml milk, and 50 ml water. Barbados sugar was used with wholemeal flour and white sugar with white flour. Approximately 0-1 mCi of 99mTc-labelled sulphur-colloid was kneaded into each bread sample which was allowed to rise for one hour and then baked at 450°C for 30 minutes. The bread was allowed to cool and was then eaten lightly buttere...