Prospective, randomized, investigator-blind trial of Helicobacter pylori infection treatment in patients with refractory duodenal ulcers. Healing and long-term relapse rates.

In this study, 26 patients with duodenal ulcers refractory to treatment with H2-receptor antagonists for 8-12 weeks were randomly assigned to eight weeks of treatment with colloidal bismuth subcitrate (120 mg four times a day) alone (N = 12) or in combination with tetracycline hydrochloride (500 mg four times a day, days 0-14) and metronidazole (500 mg three times a day, days 15-28). Symptoms were scored and endoscopy, histology, and CLO tests were performed before, on completion of treatment, and 3, 6, 12, and 18 months after treatment. Treatment was considered successful when Helicobacter pylori was not detected by CLO tests and Warthin-Starry stains on gastric biopsies taken from antrum, body, and fundus. On triple therapy, ulcers healed in 12/14 patients (85.71%) and 10/14 (71.42%) patients became Helicobacter pylori-negative. On bismuth, only one patient became Helicobacter pylori-negative (8.33%, P < 0.0001), but ulcers healed in 8/12 patients (67%, P = NS). Six patients on bismuth, whose ulcers remained unhealed or relapsed early after healing, were offered triple therapy, which resulted in ulcer healing in three and Helicobacter pylori clearance in two patients. At 18 months, none of the Helicobacter pylori-negative patients had ulcer relapse. On the contrary, ulcers relapsed in all but one patient, who remained Helicobacter pylori-positive. Smoking and drinking did not influence the therapeutic outcome. The data confirm previous reports that many duodenal ulcers are infectious and therefore curable.