Since future provider prescribing habits are established during formative training, educational interventions should be undertaken to reduce

knowledge gaps among medicine residents. W1063

AGA Abstracts

Does Low-Dose Aspirin Therapy Affect the Results of Immunochemical Fecal Occult Blood Tests? Shin Ishikawa, Jun Kato, Hideyuki Suzuki, Mitsuhiro Akita, Motoaki Kuriyama, Tsuyoshi Fujimoto, Manabu Kurome, Koji Takemoto, Toshio Uraoka, Kazuhide Yamamoto Background and aims: Low-dose aspirin is commonly used in patients with a variety of diseases such as cerebral infarction, ischemic heart disease, and atrial fibrillation. Although aspirin can cause bleeding tendency and maybe gastrointestinal bleeding, there have been few reports regarding the effect of low-dose aspirin on the results of immunochemical fecal occult blood test (iFOBT). In this study, we performed a case-control study to evaluate the performance of iFOBT on detecting colorectal neoplasia for patients with low-dose aspirin therapy. Patients and Methods: We performed a retrospective chart review analysis on 10,442 patients who underwent colonoscopy at our hospital and its 21 affiliated hospitals between October 2004 and June 2006. Among them we analyzed 233 patients (mean age 70.59.4 years; 65% male) who were regularly taking a low-dose aspirin (81 mg/day or 100 mg/day) and underwent colonoscopy for indication of a positive iFOBT result (the aspirin group). The control group was comprised of age- and gender-matched 699 subjects who underwent colonoscopy for indication of a positive iFOBT result without taking aspirin or nonsteroidal anti-inflammatory drugs. Findings of colonoscopy for these patients were investigated. Results: Among 233 patients of the aspirin group, small adenoma (adenoma 5 mm) was found in 31 (13%), advanced adenoma (adenoma 10 mm or adenoma with high-grade dysplasia) was in 29 (12%), and invasive cancer was in 5 (2%). In the control group, small adenoma was found in 115 (16%), advanced adenoma was in 99 (14%), and invasive cancer was in 16 (2%). There were no statistically significant differences in the prevalence of small adenoma (p=0.297), advanced adenoma (p=0.582), and invasive cancer (p>0.9999) between patients of the aspirin group and those of the control group. In contrast, among patients of the aspirin group, 146 (63%) patients carried no neoplasia, while in the control group, 388 (56%) patients had no neoplasia, indicating that the false positive rate of iFOBT in the aspirin group was relatively higher than that in the control group (p=0.056). Conclusions: iFOBT seems to work sufficiently for detecting colorectal neoplasia even in patients with low-dose aspirin therapy as well as in patients without the therapy, although the false positive rate was proven to be slightly higher. iFOBT should be recommended for patients with lowdose aspirin as colorectal cancer screening, although the positive predictive value may be reduced. W1064 Health Literacy Among Patients Taking NSAIDs and Proton Pump Inhibitors Eric H. Choi, Pouya Afshar, Walter Coyle Background: The proper delivery of health care is dependent on not only the clinical acumen of the physician but also the ability of patients to comply and understand the therapeutic regimen placed on them by their physicians. To elucidate the extent to which patients understand the medications that they have been placed on, many studies on health literacy have been performed, but few in the field of gastroenterology have addressed the prevalent use of NSAIDs, the inherent risk of chronic NSAID use and the proper administration of PPIs. Aim: To demonstrate (1) the prevalent use of NSAIDs in a large community-based hospital and the lack of patient education or understanding within the population to the risks of gastrointestinal bleeding and formation of ulcers as a possible complication of NSAIDs (2) the prevalent use of PPIs and the lack of patient education and compliance to the proper administration of PPIs to achieve maximal effect. Methods: An anonymous, voluntary survey was distributed to patients seen in the internal medicine, gastroenterology, and ENT offices. The survey used checkboxes to specifically detail how often patients took NSAIDs, if patients recalled their physician warning them of the risks of gastrointestinal bleeding and ulcers as a result of NSAID use, if patients were on chronic PPIs, if patients were aware of the indications for PPI administration and how patients were taking PPIs (e.g. before breakfast, with meals, randomly). Results: Of the one third of our planned 1000 patient survey, it was noted that 23.3% of patients were on daily NSAIDs, 4.7% on a weekly basis, and 54.7% on an as-needed-basis. 40% of the patients on a daily NSAID could not recall ever being told by their physician that NSAIDs can be a potential cause of gastrointestinal bleeding and ulcers. Overall, 19.8% of patients were on a daily PPI; heartburn and reflux symptoms were the reason for PPI use in 41.2% of patients, and dyspeptic symptoms (abdominal pain and distension) were the reason in an additional 23.5% of patients. It was noted that 35.3% of these patients on daily PPIs were taking their medication incorrectly (i.e. taking PPIs with meals, after meals or randomly). Conclusions: Despite the widespread use of NSAIDs and PPIs, patients continue to exhibit a poor understanding of their medications. Although it is unclear whether this lack of understanding directly stems from a true lack of patient education by physicians or an inability of patients to recall this information once presented, it highlights the need for a more systematic and concerted effort to increase health literacy in regards to NSAIDs and PPIs among patients by their physicians. W1065 Knowledge of Risk Factors for NSAID-Related Upper GI Complications and Role of Gastroprotection Among Senior Resident Physicians Gregory A. Cote, Fabiolla Siqueira, John P. Rice, Colin W. Howden Background: Despite guideline recommendations, compliance with use of gastroprotection among NSAID users is suboptimal. Previously, we have shown the combination of an electronic alert with an educational intervention modestly improves rates of gastroprotection at discharge. Our goal was to measure senior resident physician knowledge of risk factors for upper GI complications from NSAIDs. We also assessed the durability of our previous intervention. Methods: We developed a 21-question survey based on known risk factors for

upper GI complications from NSAIDs. The instrument was validated by using answers provided by 2 senior gastroenterologists not otherwise involved in the study. Cumulative scores were calculated based on answers to 16 knowledge-based questions. 5 questions explored self-perceived practice patterns. Answers were classified as consistent with or not consistent with current guidelines. The survey was administered in a monitored fashion to senior residents, some of who had received an educational intervention and a computer alert 18 months previously. We used an unpaired t-test to compare mean scores on knowledge-based questions between those residents previously exposed and unexposed to the prior intervention. Results: 37 of 38 senior residents completed the survey. 8 of 37 residents had been exposed to the intervention 18 months prior. Among knowledge-based questions, the mean score was 74% 6%. There was no significant difference in composite scores between the two groups (77 7% in the exposed vs. 73 7% in the unexposed). 57% of residents recommended a PPI for a patient requiring an NSAID who had a remote history of peptic ulcer bleeding. 57% of residents reported age > 65 or concomitant anticoagulation as very important risk factors; 24% reported a prior history of peptic ulcer as a very important risk factor, while 40% identified oral corticosteroids alone as a high risk medication. 38% of residents reported H2RAs to be an appropriate gastroprotective regimen. 73% of residents reported starting a gastroprotective regimen at discharge sometimes or often. Conclusion: Senior residents appear to be well-informed of the risk factors for upper GI complications from NSAIDs, although this has not translated into clinical practice based on their performance on clinical vignettes and prior institutional data. Awareness of age, prior ulcer disease and concomitant anticoagulation as important risk factors is suboptimal. The impact of a combined educational intervention and electronic alert did not appear to have had a durable impact on resident knowledge. W1066 Reduced Gastric Accommodation in Non-Ulcer Dyspepsia (NUD) Lasts At Least 2 Hours: Studies By Simultaneous Gastric Volume and Emptying Measurements H. Jae Kim, Michael Roarke, Michael Camilleri, Michael D. Crowell, Chris D. Tollefson, Michael S. Lee, Duane D. Burton, Alan R. Zinsmeister, Virender K. Sharma, Ananya Das Background: Approximately 40% of NUD patients have reduced accommodation in the first 30 min postprandially. It is unclear whether the accommodation response is abnormal in NUD beyond 30 min, and whether the meal volume influences the reduced gastric accommodation in health and NUD. Aims: Using simultaneous measurements of gastric volume (GV) response and gastric emptying (GE) obtained via 3-dimensional single photon emission computed tomography (SPECT [J Nucl Med 2004, Am J Physiol 2005]) we tested 2 hypotheses: 1. NUD is associated with reduced GV response to a meal and smaller differences between SPECT- measured postprandial (PP) GV and calculated intragastric volume (fasting GV plus calculated residual meal volumes at specified times) compared to health; 2. GV response is associated with maximum tolerated volume (MTV) and postchallenge (PC) symptoms in health and NUD. Methods: In 15 NUD and 16 health, validated 99m-Tc SPECT imaging using a customized analysis program (Am J Physiol 2005) was used to measure GV during fasting and GV and GE up to 3 hours (h) after the standard 111Inradiolabeled egg meal. On a separate day, MTV and PC symptoms were measured after Ensure challenge. A multiple linear regression analysis was used to assess the association of GV with NUD status adjusting for age, gender, and BMI. Similar models to predict MTV were also examined. Results: Relative to controls, NUD was associated with smaller PP volumes (PP - fasting) at all times over 2 h: at 30 min (p=0.015), 1 h (p=0.031) and 2 h (p=0.059) but not at 3 h (p=0.182). There were significant differences in measured PP GV compared to calculated intragastric volume in health at 30 min (median 718 vs. 535 ml, respectively, p=0.001) but not in NUD (637 vs.546 ml, p=NS). No significant differences were detected at 1, 2, and 3 h for both NUD and health. 2. NUD was associated with lower MTV (p<0.01) and higher PC symptoms (p= 0.017). Post-meal GV change at 1 and 2 h was a predictor of MTV (p=0.003 and 0.041, respectively, with lower volumes associated with lower MTV) and GV was a predictor of PC symptoms (higher fasting GV (p<0.001) but lower PP volumes (p<.05) at 1-3 h were both significantly associated with higher symptom scores). Conclusion: NUD is associated with impaired GV response to a meal during the first 2 PP hours. Smaller differences between measured PP GV and calculated intragastric volume in NUD (compared to health) occur in the first 30 min postprandially. Overall, smaller post-meal GV change at 1 and 2 h is associated with lower MTV and higher PC symptoms, consistent with the role of impaired gastric accommodation response over 2 hours in NUD.
W1067 Sustained Beneficial Effect of a Gluten Free Diet in Dyspeptic Patients with Lymphocytic Enteritis Without Villous Atrophy Ivar M. Harkema, Jan H. Kleibeuker, Jeroen J. Kolkman Introduction: Lymphocytic enteritis (LE) is defined as inflammation of the mucosa by intraepithelial lymphocytes without the presence of villous atrophy (Marsh I and II lesions). It can be associated with a broad spectrum of symptoms and an extensive differential diagnosis. Gluten sensitivity has been reported as possible cause of LE and a gluten free diet (GFD) could thus be beneficial. We evaluated the effect of a GFD in these patients via a questionnaire. Patients and methods: A GFD was offered to patients with LE, provided that laboratory testing, gastroscopy, and abdominal ultrasound were normal. Between 1997 and 2005 GFD was started in 135 patients. These patients were sent a questionnaire concerning abdominal and general complaints, effect of GFD and effects of gluten contamination, both before and after the GFD. We defined the success of GFD as 50 or > % improvement of symptoms. Results: 76/135 (56%) questionnaires were returned. The presenting complaints were abdominal pain (89%), diarrhoea (76%), tiredness (83%) and bloating (57%). The duration of complaints was > 1 yr in 82%. 72 % still used GFD, for mean 42 months (range 6-129 mo). 28% had stopped the diet after mean 12 months (range 1-48 mo); mostly because of non-responsiveness. Improvement of complaints after GFD was reported in 72% (in 42% between 50-75% and 30% > 75% improvement); 75% of patients reported recurrence of complaints with incidental diet violation. The time to improvement of complaints after

AGA Abstracts

A-626

W1068 Discriminant Value of Dyspepsia Subgroups According to the ROME III Consensus in Dyspeptic Patients Referred for Upper Gastrointestinal Endoscopy Joris Arts, Stefan Bourgeois, Christophe Claessens, Philip Caenepeel, Dominiek De Wulf, Lieselot Holvoet, Raf Bisschops, Gert A. Van Assche, Severine Vermeire, Jan F. Tack Background: According to the Rome 3 consensus, functional dyspepsia (FD) is subcategorized into Postprandial Distress Syndrome (PDS) and Epigastric Pain Syndrome (EPS). It has been presumed that underlying pathophysiology differs between both groups, and that acidrelated mechanisms are more likely in EPS (comprising epigastric pain and burning) than in PDS (comprising early satiation and postprandial fullness). Our aim was to investigate whether the subdivision EPS versus PDS is more likely to identify acid-related lesions in uninvestigated dyspepsia. Methods: Consecutive patients referred for open-access endoscopy for dyspeptic symptoms, who were not taking any proton pump inhibitors or H2 antagonists, were recruited. All patients filled out demographic, Rome 3 dyspepsia and heartburn wordpicture questionnaires. The symptom pattern was used to identify EPS, PDS and heartburn according to established criteria. Results: 150 patients (mean age 511.5, 65 men) participated in the study. Criteria for PDS and EPS were fulfilled by respectively 93 (62%) and 80 (53%) patients, with overlap in 61 (41%). In 32 patients (21%), the dyspeptic symptom pattern did not fulfil Rome 3 criteria for EPS or PDS. Co-existing heartburn was present in 43 patients (29%), with similar overlap with PDS (34%) and EPS (35%). Erosive esophagitis was found in 51% of the patients, mainly grade A (39%) and B (11%). A hiatal hernia, Barrett's esophagus and peptic ulcer were found in respectively 30, 6 and 8 patients, and no cancers were detected. The prevalence of erosive esophagitis was higher in patients who fulfilled criteria for EPS compared to those who did not (60% vs. 32%, p=0.003), but not for PDS versus non-PDS (43% vs. 52%, NS). The prevalence of esophagitis was also higher in those with heartburn compared to those without (69% vs. 54%, p=0.01). When patients with heartburn were eliminated, the prevalence of esophagitis in EPS versus non-EPS remained significantly different (55% vs. 29%, p=0.007), and not for the PDS group (35% vs. 53%, NS). Conclusion: When the Rome 3 criteria are applied to dyspeptic patients without anti-secretory therapy, who are referred for open-access endoscopy, major overlap between PDS and EPS exists. Compared to other groups, the EPS subgroup has a significantly enhanced prevalence of erosive esophagitis, although the predictive value is low. These observations support the presence of different pathophysiological mechanisms underlying the EPS and PDS subgroups. W1069 Role of Serine Protease Signaling in Dyspeptic Symptom Generation; a Pilot Study Ofke S. van Boxel, Jose ter Linde, Niels van Lelyveld, Peter D. Siersema, Andr Smout Background: Hypersensitivity to gastric distension has been demonstrated in a substantial portion of patients with functional dyspepsia (FD). Although mechanical hypersensitivity is considered to play a pivotal role in dyspeptic symptom generation, the molecular factors underlying this pathophysiological mechanism are unknown. Enhanced serine protease signaling via protease-activated receptor (PAR)-2 induces mechanical hypersensitivity directly by activating extrinsic primary afferents and indirectly by increasing paracellular permeability. Intestinal epithelial cells produce the serine protease trypsin IV (zymogen trypsinogen IV), which is capable of activating PAR-2. Trypsinogen mRNA expression is elevated in IBS colonic and duodenal mucosa. Moreover, IBS mucosal biopsy supernatant has been shown to induce hyperalgesia and allodynia in mice upon colorectal distension. Aim: To gain insight into the possible role of altered serine protease signaling in FD symptom generation. Duodenal biopsies were investigated, since signaling at the level of the duodenum modulates gastric sensory function. Method: Sixteen FD patients (11 female, mean age 44.8 years, range 18.9-66.5) fulfilling the Rome II criteria were included. Each patient completed a chronic dyspeptic symptom questionnaire, which graded abdominal pain or discomfort, early satiety, bloating, fullness, nausea and vomiting. Ten age- and sex-matched healthy volunteers (HV), recruited through advertisement, were studied. The presence of gastrointestinal (GI) symptoms and a history of GI disease or surgery were excluded. Any medication known to influence GI sensorimotor function was discontinued 1 week prior to upper GI endoscopy. Duodenal mucosal biopsies were taken. Trypsinogen IV transcript level was quantified by real-time RT-PCR. Gene expression was compared between patients and HV by independent student t-test. Results: No abnormalities were seen during upper endoscopy in any of the subjects. The level of Beta-actin (ACTB) expression, the reference gene used for normalization, was comparable between patients and HV. No significant difference in trypsinogen IV mRNA expression was observed between patients and HV (HV, mean 9.25, SD 3.35, range 6-16; FD, mean 8.13, SD 3.68, range 3-17; P>0.05). However, FD patients displayed a larger variability in trypsinogen IV mRNA expression. Subdivisions based on symptom pattern did, however, not reveal grouping of patients with specific characteristics. Conclusion: In contrast to IBS, trypsinogen IV mRNA expression is not increased in FD duodenal mucosa. This suggests that the serine protease trypsin IV does not contribute to dyspeptic symptom generation.

BACKGROUND: The NDI has been extensively used in clinical trials for dyspepsia and functional dyspepsia since its development approximately 10 years ago. The scale has excellent psychometric properties based on its development studies, and these have since been confirmed in clinical trials. While the responsiveness of the NDI sub-scales has been studied, the smallest change on the total or sub-scale scores that corresponds to a clinically meaningful change (MCID) has not been established. We aimed to calculate an MCID for the NDI to help guide interpretation of future trials. METHODS: Comprehensive clinical data was collected in outpatients (n=101) and community subjects (n=460), and the MCID for the NDI total score was examined in three ways. The first estimates the average change (over a 2 week period) in the NDI corresponding to a 5-point change in the SF-36 mental and physical sub-scales, the second repeats this using a one-grade change in symptom level, while the third is to calculate Cohen's d effect size measure among individuals changing by at least 5 points on the SF-36 sub-scales. A separate cross-sectional study was used to obtain the receiver-operator characteristic (ROC) curve for discriminating between dyspepsia and non-dyspepsia subjects. RESULTS: Among individuals changing by at least 5 points on the SF-36 physical scale, the corresponding median change in the NDI total score was 12 points (SD=12 points), and for the mental scale the median NDI change was also 12 points (SD= 9 points). Results were similar for individuals changing by at least one grade on a 5 point global symptom Likert scale (median NDI change 11, SD = 11). Based on comparing individuals changing by at least 5 points on the SF-36 physical sub-scale with those who did not Cohen's d was calculated to be 0.25, considered to be in the moderate range of effect sizes. In a separate cross-sectional community study, the NDI QoL total score was shown to provide good discrimination between individuals meeting and not meeting the Rome criteria for functional dyspepsia, with an area under the ROC curve of 0.84 (95 % confidence interval 0.80, 0.89). CONCLUSIONS: A change of at least 10 points on the NDI total scale is suggested to correspond to a clinically meaningful change in patient status, since this corresponds to an approximately 5 point change in general quality-of-life and a 20% change in globally assessed symptom score. Based on clinical trial data, a 10-point change also corresponds to approximately 1 SD and 10% of the total scale range, both of which are intuitively clinically relevant changes. W1071 Novel Associations with Dyspepsia: A Community Based Study Nicola Gathaiya, E. Janet Castillo, G. Richard Locke, Michael Camilleri, Cathy Schleck, Alan R. Zinsmeister, Nicholas J. Talley Introduction. Dyspepsia is a common condition but the risk factors identified to date have not improved the care of patients. We evaluated novel associations with dyspepsia, specifically family history and sleep disorders. Methods. An age (5 year interval) and sex (equal number of men and women) stratified random sample of Olmsted county residents aged 20-80 years was drawn from a list of respondents to previous studies. A modified version of the validated bowel disease questionnaire (BDQ) which assessed both GI symptoms and novel associations was sent to 484 individuals, of which 224 responded (46%). Dyspepsia was defined as having troublesome upper abdominal, stomach or belly symptoms for at least 25% of the time in the last year. Multiple variable logistic regression was used to estimate the odds ratio (ORs) [95% confidence intervals (CI)] adjusting for age, sex, BMI and acid suppression medications. Results. A total of 48 (22%) subjects reported symptoms consistent with dyspepsia. Five significant associations were identified. Among those who reported a positive family history of indigestion, 37% reported dyspepsia, (OR=3.4, 95% CI (1.0-11.5), p= 0.048), and among those with a positive family history of abdominal pain, 45% reported dyspepsia, (OR=4.7, 95% CI (1.5-14.8), p=0.008). Among those reporting difficulty falling asleep, 38% had dyspepsia, (OR=8.2, 95% CI (2.2-31.5), p=0.002), and in those reporting poor sleep associated with worsening symptoms, 79% had dyspepsia, (OR=15.9, 95% CI (2.0-124.9), p=0.009). Finally, in those reporting high somatic symptom check-list (SSC) score, 34% reported dyspepsia, (OR=5.6, 95% CI (1.5-20.6), p=0.01). Family history of GERD and bowel problems, race, marital status, alcohol and nicotine use were not significantly associated with dyspepsia. Conclusion. Familial aggregation raises the possibility of a genetic component although shared environmental factors need to be considered. Sleep dysfunction and somatization suggests a psychological component. These data provide new avenues for approaching this difficult condition. W1073 Analysis of the Predictive Value of Alarm Symptoms in the Diagnosis of Gastric Malignancy in the Presence of Gastric Ulceration Ophelia K. Veraitch, Constantine Alifrangis, Mohamed Shariff, Paul M. Trembling, Anil Goburdhone, Tahreema N. Matin, Rakesh Chaudhary, John P. Martin, Geoff Smith, Andrew V. Thillainayagam, Devinder S. Bansi Introduction To maintain appropriate referral for urgent gastroscopy, alarm symptoms are used to select patients thought to be at higher risk of significant upper gastrointestinal pathology. In the UK, the National Institute of Clinical Excellence (NICE) classify alarm symptoms as gastrointestinal bleeding, dysphagia, unintentional weight loss, abdominal swelling and persistent vomiting. These guidelines are employed to differentiate simple dyspepsia from upper gastrointestinal malignancy and peptic ulcer disease. Aim To assess the predictive value of alarm symptoms in diagnosing gastric malignancy in gastric ulcer patients. Methods We performed a retrospective case note analysis of all patients diagnosed with gastric ulcers between 1999 and 2007 at a large teaching hospital in the UK, using the in-house electronic endoscopy database (Endoscribe). The position, size and endoscopic appearance of the ulcer (rolled edge, flat edge, necrotic base etc.) was noted. Histopathology reports were analysed for all ulcers that were biopsied. We determined the proportion

A-627

AGA Abstracts

AGA Abstracts

start of the GFD was 15% at 1 mo, 58% at 3 mo, 83% at 6 mo, and 97% at 12 mo. The time to an optimal response of the GFD was 3% at 1 mo, 25% at 3 mo, 45% at 6 mo, and 78% at 12 mo. A significant correlation was found between the improvement on GFD and the severity of complaints after incidental diet violation (r = 0,24; p = 0,04). In conclusion: Gluten seems to be a causative factor in a large proportion of patients with LE and otherwise unexplained dyspepsia. The response of symptoms to GFD in these patients is slow. Therefore, a GFD should be considered in patients with LE without villous atrophy for a period of at least 6 months.

W1070 Estimation of the Minimum Clinically Important Difference (Mcid) for the Nepean Dyspepsia Index (Ndi), a Validated Quality of Life (QOL) Scale for Functional Dyspepsia Mike Jones, Nicholas J. Talley