51.   Table IV shows the prevalence of active duodenitis and gastric metaplasia in the presence of absence of the various types of gastritis.

52.   Logistic regression showed that gastric H pylori and the combination of active duodenitis and gastric metaplasia were independent predictors of duodenal ulceration.

53.   Table V shows the distribution of NSAIDs, with or without second line drugs in patients with active duodenitis and gastric metaplasia.

54.   Patients taking diclofenac, ketoprofen or nabumetone were less likely to have active duodenitis and gastric metaplasia than other NSAIDs but the differences were not statistically significant.

55.   The study also shows a strong association between duodenal ulcers, active duodenitis, gastric metaplasia, and H pylori related gastritis regardless of NSAID intake.

56.   Gastric metaplasia tended to occur less frequently in our patients receiving NSAIDs, but the differences did not reach statistical significance.

57.   High acidity of the duodenal contents has for a long time been found to be associated with gastric metaplasia, both in humans and in laboratory animals.

58.   Acute exposure to NSAIDs is also known to stimulate gastric acid secretion, which, at least in theory, should increase the prevalence of gastric metaplasia.

59.   Similar numbers of duodenal ulcers were also seen in patients with active duodenitis or gastric metaplasia, or both, regardless of NSAID intake.

60.   Most patients with duodenal ulcers, however, in the presence or absence of NSAIDs, were found to have active duodenitis and gastric metaplasia.