A182 AGAABSTRACTS • G0744 • G0746 ENDOTHELIN, RELEASED FROM THE FUNDUS OF THE STOMACH, PARTICIPATES IN THE REGULATION OF ACID SECRETION. Shashi Kodnru, Ari Hirsch, Robert McCuen, and Mitchell L. Schubert. Medical College of Virginia/McGuire VAMC, Richmond, VA. Endothelins comprise a family of 21-amino acid peptides that were initially shown to mediate vasoconstriction; three isoforms, ET-1, ET-2, and ET-3, and two receptor subtypes, ETA and ETa, have been identified. Recently, ET-I was localized by immunohistochemistry to rat parietal ceils (Gastro 102:A156, 1995). In the present study, we have used selective endothelin receptor antagonists to examine the role of endogenous ET- 1 in the regulation of acid secretion. The isolated mouse stomach was luminally perfused with saline and acid secretion measured by titration. Addition of the ETa antagonist, BQ-788 (0.1 pM), for 20-min caused an increase in acid secretion (11 ± 4%; P < 0.05), implying that endogenous endothelin, acting via ETa receptors, inhibits acid secretion. Consistent with this notion, addition of ET-3 (1 pM to 1 ktM), which interacts preferentially with ETa receptors, caused a concentration-dependent decrease in acid secretion. The maximal response occurred at 10 nM (34 ± 3% below basal level; P < 0.01) and the IC5o was 0.2 nM. In contrast, the ETA antagonist, BQ-485 (0.1 pM), had an opposite effect, slightly decreasing acid secretion (5 -+ 1% below basal level; P < 0.05), implying that endogenous endothelin, acting via ETA receptors, stimulates acid secretion. The ability of endogenous ET-1 to interact with both ETa and ETA receptors is exemplified by the effect of exogenous ET-1 on acid secretion. ET-1 (1 pM to 1 pM) caused a concentration-dependent decrease in acid secretion. The maximal response occurred at 100 nM (33 ± 1% below basal level; P < 0.01) and the IC50 was 5 nM. Preincubation with the ETa antagonist converted the decrease in acid secretion elicited by ET-I (10 nM) to a significant increase above basal level (24 ± 3%; P < 0.01). Preincubation with the ETA antagonist augmented the decrease in acid secretion induced by ET-1 from 17 ± 2% to 24 ± 1% below basal level (P < 0.05 for the difference). A combination of ETA and ETa antagonists abolished the acid response to ET-1, restoring it to basal level. We conclude that ET-1, released from parietal cells, exerts dual inhibitory and stimulatory influences on acid secretion. The dominant effect is inhibitory and mediated via the ETa receptor; a coexistent stimulatory effect, mediated via the ETA receptor, is unmasked when the inhibitory pathway is blocked. • G0745 CONTRIBUTION OF ACID AND BILE REFLUX TO SYMPTOMS IN • PATIENTS WITH GASTRO-ESOPHAGEAL REFLUX DISEASE. G. Koek, J. Tack, A. Degreef, T. Lerut, J. Janssens. Center for Gastroentero- logical Research, K.U. Leuven, Belgium. Recent studies, using optical monitoring of esophageal bilirubin concentration, have suggested that mixed reflux of both acid and duodenal contents frequently occurs in patients with gastro-esophageal reflux disease. Only one study assessed the role of duodeno-gastro-esophageal reflux in the genesis of esophageal symptoms (Marshall et al., 1997). However, in that study, 45% of the symptom episodes (occurring during the meal or postprandially) could not be studied because of the use of a solid meal. Aim: To study the contribution of acid and bile reflux to symptoms in patients with presumed reflux disease, including the meal and postprandial period. Methods: Forty two patients (23 men, mean age 43 ± 3) with presumed gastro-esophageal reflux disease underwent 24 hour ambulatory pH and Bilitec monitoring. During the study, only a liquid diet not interfering with Bilitec monitoring was used, and patients were instructed to press a marker button to record symptom events. For each symptom episode, the minimal pH and the maximal bilirubin absorbance in a 4 minute window centered around the event marker were assessed. For each patient, a symptom index was calculated in relation to both acid and bile reflux episodes. Results (mean ± SEM) were compared by t-test. Results: A total of 273 symptom episodes were identified. Sixty five symptom episodes (24%) were associated with acid reflux alone (pH<4), 13 (5%) with bile reflux alone (absorbance > 0.14) and 28 (10%) with both. No acid or bile reflux was present for 167 symptom episodes (61%). Pathological esophageal acid exposure ( > 4% of the time pH below 4) was present in 24 patients. When the analysis was limited to this subgroup, out of 117 symptom episodes, 33% were associated with acid reflux alone, 4% with bile reflux alone and 23% with both. For the whole group of 42 patients, a positive symptom index (>-75%) for acid reflux was present in 15 patients (36%) and for bile reflux in 7 patients (17%). However, all patients with a positive symptom index for bile reflux had also a positive symptom index for acid reflux. Conclusions: Even including the meal and postprandial period, symptom episodes in patients with presumed gastro-esophageal reflux disease are much more often related to acid reflux than to bile reflux. Duodeno-gastro-esophageal reflux does not seem to be important in producing esophageal symptoms. GASTROENTEROLOGY rot. 114, No. 4 EFFECT OF ERYTHROMYCIN ON GASTRIC EMPTYING AND MEAL-RELATED SYMPTOMS IN FUNCTIONAL DYSPEPSIA. G. Koek, J. Tack, B. Geypens, Y. Ghoos, J. Janssens. Dept. of Gastroenterology, U.Z. Gasthuisberg, K.U. Leuven, Belgium. Current pharmacological treatment of functional dyspepsia focuses on the use of gastroprokinetic drugs. Erythromycin is a powerful prokinetic through its action as a motilin receptor agonist. However, the therapeutic potential of erythromycin in functional dyspepsia has not been thoroughly evaluated. Aim: To study the effect of erythromycin on gastric emptying rate and on meal- related symptoms in patients with functional dyspepsia. Methods: In thirteen patients with functional dyspepsia (9 women and 4 men, mean age 48 ± 5 years) gastric emptying was measured twice, after administration of placebo or erythromycin 200 mg i.v. in a double-blind fashion. Gastric emptying rates for solids and liquids were determined using the 14C octanoic acid and 13C glycin breath test. Breath samples were taken before the meal and at 15 minute intervals for a period of 240 minutes postprandially, and gastric half emPtYing time (tl/2) was calculated from the data. At each breath sampling, the patient was asked to grade the intensity (0-3; 0--absent, l=mild, 2=relevant and 3=severe) of six different symptoms (epigastfic pain, bloating, postprandial fullness, nausea, belching and epigastric burning). For each symptom, a meal-related severity score was obtained by adding all intensities over the whole study period. A cumulative symptom score was obtained by adding individual symptom severity scores. All data are given as mean-+ SEM. They were compared using paired Student's t test. Results: Prior treatment with erythromycin enhanced the gastric emptying rate for solids (tl/2 58.4 ± 6.1 vs. 93.6 ± 16.4 min. after placebo, p < 0.05) and for liquids (71.6± 17.9 vs. 86.6±2.52 min. after placebo, p<0.01). Erythromycin had no effect on the cumulative meal-related symptom score (66.0 ±7.7 vs. 89.8 ± 17.9 min. after placebo, NS). Except for bloating (11.8 ± 3.7 vs. 17.9 ± 4.6 min. after placebo, p < 0.05), individual symptom severity scores were not altered by erythromycin (epigastric pain 10.5 ± 4.9 to 15.0 ± 4.3, postprandial fullness 18.8 ± 5.8 to 14.3 ± 5.2, nausea 12.3 ± 6.5 to 9.5 ± 4.3, belching 18.5 ± 4.3 to 14.8 ± 3.7 and epigastric burning 8.6 ± 3.6 to 6.3 ± 2.5, all NS): Conclusion: In spite of its gastroprokinetic effect, erythromycin falls to improve meal-related symptoms in patients with functional dyspepsia. These findings are pleading against impaired gastric emptying as the sole mechanism underlying symptoms in functional dyspepsia. G0747 TREATMENT OF HELICOBACTER PYLORI (liP) DOES NOT IMPROVE SYMPTOMS OF FUNCTIONAL DYSPEPSIA (FD). H.R. Koelz, Zurich, Switzerland; R. Arnold, Marburg, Germany; M. Stolte, Bayreuth, Germany; A.L. Blum, Lausanne, Switzerland; and FROSCH Study Group. Background and aim: The role of HP in FD is controversial. We investigated the effect of HP treatment on symptoms of FD. Methods: This double-blind randomized controlled trial was carried out in Germany. Included were HP-positive patients (gastroscopy with rapid urease test and 13urea breath test, UBT) with epigastric symptoms who had not responded satisfactorily to a 2-week double-blind acid-inhibitory treatment (Gastro 1997; 113:A73). We excluded patients with heartburn/acid regurgitation without concomitant epigastric symptoms and/or predominant irritable bowel symptoms as well as those with abnormal abdominal sonography or gastroscopy other than HP gastritis and/or < 10 gastric erosions. The patients were treated for 2 weeks with omeprazole 40mg bid and amoxicillin lg bid (OME+AMO) or omeprazole 20mg om alone (OME). UBT was repeated 6 weeks after the beginning of treatment. Follow-up was 6 months. The primary outcome variable (response) was defined as "no need for further therapy or investigations 4 to 6 months after treatment, according to the investigator's judgment". Results: Randomized were 181 patients (mean age 47 ± 15 SD years, 41% males). In an Intention-To-Treat-Analysis, the response rate was high, but similar in the two treatment groups and also similar in patients who became HP-negative as compared to those who remained HP-positive (Table). Patients with missing values had left the study early because of particularly poor response. Responders after 6 months HP (UBT after 6 weeks) OME OME+AMO Total Negative 8/9 (89%) 33/46 (72%) 41155 (75%) Positive 51/69 (74%) 20•35 (57%) 711104 (68%) Missing 2/14 (14%) 2/8 (25%) 4•22 (18%) Total 61192 (66%) 55/89 (62%) 116/181 (64%) Similar results were found in a Per-Protocol-Analysis. Furthermore, the groups, when compared according to type of treatment and healing of HP infection, showed no significant differences in total symptom scores (visual analogue scale), fractions of patients who became completely asymptomatic, individual FD symptoms, and quality of life measures. In a logistic regression analysis using "response" as the dependent variable, no baseline characteristic could be identified that significantly correlated with outcome.