AGAAbstracts
`
`558
`
`Colorectal Cancer Screening in Chinese Immigrants: Who Can Deliver the
`Message?
`Karen E. Kim, Stacy Tong, Eva Lu-Bonn
`
`Background: Colorectal cancer (CRC) ranks third in cancer incidence and mortality in Asian
`Americans. Studies on colorectal cancer in Asian Americans suggest that screening rates are
`among the lowest reported. Barriers to colorectal cancer screening (CRS) in immigrants
`include language, access and absence of routine health care. The effectiveness of ethnic and
`culturally specific community outreach on colorectal screening in Chinese immigrants has
`not been previously reported. This study compares the effectiveness of physician directed
`compared to lay health educator directed CRS programs in Chinese immigrant communities.
`Methods: Chinese community based organizations were asked to randomize their clients
`aged 50 and above into two CRS programs. Physician directed programs used Asian, non-
`Chinese speaking physicians coupled with a trained medical interpreter for CRC education
`and fecal occult blood testing (FOBT), while lay educator programs (LHE) used peer educators
`for education on CRC and FOBT. All participants completed a translated survey on CRS
`compliance, CRC risk perception and knowledge gained. The effectiveness of each educational
`program on CRS behavior was assessed by compliance with fecal occult blood testing (FOBT).
`Results: 260 participants attended CRC programs. The average age was 54; 55% were women,
`the average years in the US was 10. 80% of participants reported speaking Chinese at home
`and were of limited English proficiency. 5% of participants in the MD group (n=135)
`completed the FOBT, compared to 76% in the LHE group (n=125). Overall, 30% were
`found to have a positive FOBT. 75% in the MD group and 85% in the LHE group reported
`a gain in knowledge about CRC and CRS. Conclusions: In this study, Chinese immigrants
`were significantly less likely to comply with FOBT when education was delivered by Asian,
`non-Chinese speaking physicians. In stark contrast, the same program delivered by a peer
`educator (a respected member of their community) resulted in an extremely high FOBT
`compliance. Since no significant difference in knowledge gained between the two groups
`were reported, these differences cannot be attributed to language barriers. Given the high
`rates of positive FOBT in this population, this study highlights the importance of community
`based and community directed education in hard to reach populations.
`
`559
`
`Variation in Polyp Detection Rates At Screening Colonoscopy
`Thomas F. Imperiale, Elizabeth A. Glowinski, Ching L. Cooper, Gregory N. Larkin,
`Faouzi Azzouz, David F. Ransohoff
`
`Background: Wide variation in adenoma detection rates has been found in the U.K. Flexible
`Sigmoidoscopy Screening Trial. This variation, attributed to endoscopist performance, has
`been used to justify the need for establishing quality standards for performance of colonos-
`copy. Purpose: The objectives of this study were to measure, in a program of screening
`colonoscopy, variation in polyp detection rates (PDRs) and to identify factors associated
`with PDRs. Methods: We reviewed summary-level data from endoscopists who performed
`at least 40 first time screening colonoscopies for the Lilly Colorectal Cancer Prevention
`Program between 9/1995 and 6/2001. Summary-level data include mean age, mean procedure
`time (MPT), gender distribution, and % persons with any polyp, any adenoma, any polyp
`>= 1.0 cm, and multiple polyps. We described the variation in PDRs and used boxplot
`analysis to identify outliers. Multivariable regression modeling identified factors associated
`with PDRs and that accounted for some of the variability in PDRs. Results: 2664 screening
`colonoscopies (1108 [42%] in women, 1556 in men) were performed by 25 endoscopists,
`whose number of procedures ranged from 40 to 207. Overall mean patient age was 59 years
`(mean age range, 56 to 62 yr), the mean % women was 42% (range, 29-52%), and MPT
`was 17.1 minutes (range, 9.7-31 min). Mean PDRs, ranges, number of outliers, and outlier
`values are in the Table. Regression models that included procedure N, mean age, % women,
`and MPT accounted for 42% of the variation in PDRs for any polyp, 56% of the variation
`in PDRs for any adenoma, 37% for large polyps, 36% for > 1 polyp, and 46% for >1 adenoma.
`In all models, only MPT was significantly associated with PDRs. For the endoscopists with
`the shortest (9.7 minutes; N=80) vs. longest (31 min; N=56) MPTs, respective PDRs for
`any polyp, any adenoma, polyps >= 1 cm, > 1 polyp, and > 1 adenoma were 20% vs. 82%,
`15% vs. 43%, 0% vs. 12.5%, 7.5% vs. 50%, and 6.3% vs. 25% (all P-values < 0.003).
`Conclusion: PDRs vary widely among screening colonoscopists, though only a few (high)
`outliers were identified. Variation in PDRs was significantly associated with MPT, but not
`with age, gender, or the number of procedures. Further research is needed to understand
`the reasons for variation, and requires attention to patient risk factors, scope withdrawal
`time, and quality of the prep.
`
`560
`
`The Impact of Literacy On Colonoscopy
`Rony Ghaoui, Arthur T. Evans, Bashar M. Attar
`
`Background and Aims Approximately 20-30% of attempted colonoscopies are incomplete
`because of poor bowel preparation. This problem is not only costly and wasteful, but
`increases risks and delays diagnosis and treatment. Although literacy skills are necessary for
`patients to successfully follow home bowel preparation instructions, they are nearly imposs-
`ible to assess informally. Our aim was to evaluate the impact of literacy on bowel preparation
`and completion of colonoscopy, with the ultimate objective of redesigning instructions to
`be more effective, regardless of patients’ literacy skills. Methods We performed a prospective
`
`T : 89386$$CH2
`05-04-06 23:18:25
`Layout: 89386B : o
`
`Page 81
`
`cohort study at a large urban public hospital during the summer of 2005 among patients
`who presented for outpatient colonoscopy. Before the colonoscopy, literacy was assessed
`using the 7-minute test of functional health literacy. Using structured interviews, we also
`gathered information on patients’ past history, beliefs, and self-reported behaviors on bowel
`preparation. Results of the colonoscopy were obtained from the procedure reports. Results
`The sample of 195 subjects included 64% women, 49% African Americans, and 40% with
`low literacy skills. 25% of colonoscopies were not completed, primarily (90%) because of
`poor bowel preparation. Low literacy increased the risk of an incomplete exam 4-fold: 12
`% incomplete exams among patients with adequate literacy skills compared to 45% with
`low literacy skills (p<0.001). Other important predictors of an incomplete exam were: eating
`lunch or dinner the previous day, not taking the bisacodyl (p=0.001), and not finishing the
`1-gallon polyethylene glycol solution (p=0.01). Patients with a past history of colonoscopy
`(p=0.03) and those who received additional oral instructions from a physician or nurse (p=
`0.05) were more likely to have a complete colonoscopy. Low literacy remained the strongest
`independent predictor of incomplete colonoscopy after controlling for the contribution of
`all other factors (OR=8.7; 95% CI: 3.4-22.1; p<0.001). Discussion/Conclusion Low literacy
`accounts for 30% of all incomplete colonoscopies. This important effect is independent of
`age, gender, ethnicity, and even whether oral instructions supplemented the written instruc-
`tions. In light of these findings, we must redesign our methods of instructing patients about
`bowel preparation in order to improve the efficiency and effectiveness of colonoscopy.
`
`561
`
`Prevention of Low Dose Aspirin-Associated Gastroduodenal Ulcers and Upper
`Gastrointestinal Symptoms in Patients Receiving Esomeprazole 20 Mg Per Day
`Neville D. Yeomans, Angel Lanas, Joachim Labenz, Sander Veldhuyzen van Zanten,
`Christo van Rensburg, Istvan Racz, Emma Naucler, Lars-Erik Svedberg
`
`OBJECTIVE: To assess whether esomeprazole reduces the incidence of ulcers in patients
`taking low dose (75-325 mg daily) aspirin (acetylsalicylic acid). METHODS: Male or female
`Helicobacter pylori-negative patients ≥60 years with a condition requiring daily low dose
`aspirin, at risk of developing gastroduodenal ulcers were included in this randomized,
`double-blind, multicenter, placebo-controlled trial. Patients were excluded if they had upper
`gastrointestinal symptoms, including heartburn, that required treatment, or if they had
`erosive esophagitis. Patients with Los Angeles Grade A lesions and without reflux symptoms
`requiring treatment could be included. Patients were randomized to receive either esomepra-
`zole 20 mg or placebo once daily (qd) for 26 weeks. Patients continued to receive the aspirin
`dose they were taking when enrolled into the study. The primary outcome variable was the
`presence of gastric and/or duodenal ulcers at endoscopy over the 26-week period. We
`present data from endoscopy taken at 26 weeks. RESULTS: A total of 991 patients (57.1%
`male, mean age 69.3 years, median aspirin dose 100.0 mg, 89.0% taking aspirin for >4
`weeks) were included in the intent-to-treat population. Eight patients (1.6%) had developed
`a gastroduodenal ulcer in the esomeprazole group by 6 months, compared with 27 patients
`(5.4%) in the placebo group (p=0.0007). This corresponded to a relative reduction of
`developing an ulcer of 70% when taking esomeprazole rather than placebo. The related life
`table estimates were 1.8% for esomeprazole and 6.2% for placebo. In patients with Grade
`A esophageal lesions at baseline and with pre- and post-baseline data, 13 patients (28.3%)
`in the esomeprazole group had Grade A lesions at week 26, compared with 37 patients
`(71.2%) in the placebo group (p<0.0001). Resolution of investigator-assessed aspirin-associ-
`ated upper gastrointestinal symptoms (rated as none to severe) was significantly higher with
`esomeprazole than with placebo for epigastric pain, burning and discomfort, as well as
`heartburn and bloating (p<0.05 for all symptoms). Esomeprazole was safe and well-tolerated.
`CONCLUSIONS: Esomeprazole 20 mg qd is more effective than placebo and is well-tolerated
`in reducing the risk of developing gastroduodenal ulcers and resolving upper gastrointestinal
`symptoms associated with low dose aspirin use in patients at moderate-to-high risk of
`developing ulcers. At risk patients taking low dose aspirin may therefore benefit from co-
`administration of esomeprazole.
`
`562
`
`Effects of Concomitant Aspirin (81 mg qd) On Incidence of Gastric and/Or
`Duodenal Ulcers in Healthy Subjects Taking Celecoxib Or Naproxen: A
`Randomized, Placebo-Controlled Trial
`Jay L. Goldstein, James Aisenberg, Manuela Berger, William E. Dodge
`
`Background: Patients who use nonselective nonsteroidal anti-inflammatory drugs (NSAIDs)
`or cyclooxygenase-2 (COX-2) selective inhibitors often require concomitant low-dose aspirin
`(ASA) for cardiovascular (CV) prophylaxis. The impact of the most common prophylactic
`ASA dose (81 mg qd) on the gastrointestinal (GI) safety profile of COX-2 selective inhibitors
`remains controversial. Methods: In this multicenter, double-blind, placebo-controlled study,
`healthy subjects (seronegative for Helicobacter pylori) aged 50-75 y were randomized to
`receive celecoxib 200 mg qd + ASA 81mg qd, naproxen 500 mg bid + ASA 81 mg qd, or
`placebo + ASA 81 mg qd for 1 week. The primary end point was incidence of gastric and/
`or duodenal ulcers in subjects treated with celecoxib + ASA compared with naproxen +
`ASA. Upper GI (UGI) endoscopy was performed at baseline and Day 7 (or early termination).
`Subjects were excluded if baseline endoscopy revealed >5 erosions in the stomach or
`duodenum; any gastric, pyloric channel, or duodenal ulcer (≥3 mm diameter); or any
`esophageal ulcers/erosions. In addition, subjects with any NSAID use within 2 weeks of
`enrollment were excluded. Results: A total of 662 subjects (mean age, 58 y) were randomized
`in a 2:2:1 fashion (celecoxib + ASA, n=267; naproxen + ASA, n=264; placebo + ASA, n=
`131). There were no significant differences in baseline demographics between groups. A
`total of 7% (18/257) of evaluable subjects randomized to celecoxib had gastric and/or
`duodenal ulcers compared with 25.3% (65/257) of those on naproxen and 1.6% (2/129)
`of those on placebo. Statistical pairwise comparisons by Cochran-Mantel-Haenszel test are
`shown in the Table. For secondary end points, significantly fewer celecoxib-treated subjects
`had gastric ulcers (GU) or duodenal ulcers (DU) vs naproxen (GU: 5.8% vs 22.6%; DU:
`1.2% vs 7%; P<0.001 both comparisons). More celecoxib-treated subjects had GUs vs
`placebo (P=0.016) but there was no significant difference in the incidence of DUs (1.2% vs
`0.8%). Conclusion: In a healthy population taking 81 mg qd ASA, celecoxib co-administration
`
`A-81
`
`AGA Abstracts
`
`Page 1 of 2
`
`Patent Owner Ex. 2017
`CFAD v. Pozen
`IPR2015-01718
`
`

`
`Gastroprotection (GP) gap in elderly patients 1995-2005. PPI=proton-pump inhibitors,
`miso=misoprostol, cox-2=selective cox-2 inhibitors
`
`565
`
`Improved Gastrointestinal Safety Profile with Lumiracoxib Compared with
`Naproxen and Ibuprofen in Patients At Least 65 Years Old At Increased Risk
`of Gastrointestinal Events
`Christopher Hawkey, Wilfred Weinstein, Walter Smalley, Gerhard Krammer, Bernhard
`Mellein, Dominik Richard, Xavier Gitton
`
`Background: Increasing age is a risk factor for nonsteroidal anti-inflammatory drug (NSAID)-
`related gastrointestinal (GI) ulcers and associated ulcer complications. Selective COX-2
`inhibitors such as lumiracoxib were developed to reduce the risk of GI events. The Therapeutic
`Arthritis Research and Gastrointestinal Event Trial (TARGET), the largest published GI
`outcomes study to date, showed a 79% reduction in GI ulcer complications (perforation,
`obstruction or bleeding) for lumiracoxib compared with naproxen and ibuprofen in the
`non-aspirin population. Here we report additional analyses on the sub-group of TARGET
`patients ≥65 years of age. Methods: Patients (n=18 325) aged ≥50 years with OA were
`randomized to receive lumiracoxib 400 mg once daily (od) (4 times the recommended dose
`for treatment of OA), naproxen 500 mg twice daily (bid) or ibuprofen 800 mg three times
`daily (tid) for 1 year. Randomization was stratified by age (50-64, 65-74, ≥75 years)
`and low-dose aspirin use. The primary endpoint was definite or probable upper GI ulcer
`complications. Secondary GI endpoints included all definite or probable ulcers (complicated
`and uncomplicated symptomatic ulcers) and clinically evident major bleeds. Results: In
`total, 7939 of the TARGET population were ≥65 years (lumiracoxib [n=3980] and NSAIDs
`[n=3959]), approximately 31% of whom took low-dose aspirin. In patients ≥65 years not
`taking aspirin (n=5484), there was a significant reduction in ulcer complications in the
`lumiracoxib group compared with the NSAID group (0.29% vs 1.42% respectively; p<0.0001)
`and significantly fewer ulcers (complicated and uncomplicated) (0.69% vs 2.62%; p<0.0001).
`In addition the incidence of clinically evident major bleeds was significantly lower in the
`lumiracoxib group than in the NSAID group (0.26% vs 0.98%; p=0.0012). In patients ≥65
`taking low-dose aspirin, there were no statistically significant differences between treatments.
`Conclusion: Lumiracoxib 400 mg (od) delivers a consistent GI safety profile that is superior
`to the NSAIDs naproxen or ibuprofen in patients ≥65 years not taking aspirin.
`
`566
`
`A Prospective, Randomized, Controlled Trial of Clear Liquids Vs. Low-Fat
`Solid Diet as the Initial Meal After Mild Pancreatitis
`Brian C. Jacobson, Martha B. Vandervliet, Michael Hughes, Rie Maurer, Katherine D.
`McManus, Peter A. Banks
`
`Background: Patients recovering from mild acute pancreatitis typically receive a clear liquid
`diet (CLD) when ready to initiate oral nutrition. The timing of patient discharge frequently
`depends upon successful advancement to solid food. We hypothesized that initiating feeding
`with a low-fat solid diet (LFSD) after mild pancreatitis would be well tolerated and would
`result in a shorter hospital length of stay (LOS). Methods: A sample size of 120 subjects
`was calculated to have 90% power to detect a one day difference in LOS between two study
`arms. Exclusion criteria included narcotic use within 6 hours prior to refeeding, an underlying
`condition that could itself cause poor oral intake or prolonged LOS, inability to monitor
`the patient post-discharge, pregnancy, and supervision of a patient’s care by a study team
`member. Patients with mild pancreatitis, defined in accordance with the Atlanta symposium,
`were randomized to receive either a CLD or a LFSD when the responsible medical team
`determined refeeding was appropriate. The decision to advance a patient’s diet and the
`timing of discharge were determined by the medical team. Patients were monitored daily
`for dietary intake, recurrence of pain, need to stop feeding, post-refeeding LOS, and for a
`total of 28 days post-refeeding to capture symptoms or readmission. Post-refeeding LOS
`(primary outcome) was compared using a t test and need to stop feeding (secondary outcome)
`was compared with Fisher’s Exact test. Results: We randomized 121 patients: 66 to CLD,
`55 to LFSD. Baseline characteristics including age, gender, and cause of pancreatitis were
`similar in both groups. The number of patients requiring cessation of feeding because of
`pain, nausea or vomiting was similar in both groups (5% for CLD and 9% for LFSD; p=
`0.47). There was no difference in 28 day re-admission rates between the two arms. By
`intention-to-treat, the mean LOS after refeeding was similar in both groups (1.7 +/- 1.9
`days for CLD and 1.7 +/- 2.0 days for LFSD; p=0.94). Calorie counts for the first meal
`consumed were available for 69% of patients and showed significantly lower calories and
`grams of fat consumed in the CLD arm than in the LFSD arm (mean 177 +/- 106 cal vs
`394 +/- 250 cal; 2 +/- 4 gm fat vs 8 +/- 7 gm fat; p<0.001 for both comparisons). Per-
`protocol analyses and analyses restricted to those with calorie counts provided similar results
`for both diet tolerability and post-refeeding LOS. Conclusions: Initiating feeding with a
`LFSD was as well tolerated as a CLD, but did not result in a shorter LOS.
`
`results in more gastric and/or duodenal ulcers than aspirin alone but fewer compared with
`naproxen. In parallel with a similarly designed trial (but with 325 mg qd ASA), these data
`suggest that daily ASA doses commonly used for CV prophylaxis do not completely negate
`the UGI benefit of celecoxib compared with naproxen, as measured by endoscopic ulcer rates.
`
`563
`
`Effects of Antisecretory Drugs and Nitrates On the Risk of Ulcer Bleeding
`Associated with NSAIDs and Anti-Platelet Agents
`Angel Lanas, Luis A. Garcia-Rodriguez, Maria T. Arroyo, Fernando Gomollon, Faust Feu,
`Montse Forne, Sofia Aleman, Enrique Garcia, Luis Bujanda, David Nicolas, On Behalf of
`AEG Investigators
`
`Background and Aims: After the withdrawal of rofecoxib, an increased prescription rate of
`some non-selective NSAIDs has been observed, but, according to recent reports, additional
`prevention strategies are not being followed. In this study we report the effect of antisectory
`drugs (proton pump inhibitors-PPI; H2-receptors antagonists- H2-RA) and nitrates on the
`risk of upper gastrointestinal ulcer bleeding (UGIB) associated with NSAID use in clinical
`practice Methods: Type of Study: hospital-based case-control study with prospective data
`collection. Setting: A network of 40 general hospitals integrated within the Spanish Associ-
`ation of Gastroenterology. Cases were consecutive patients with UGIB confirmed by endos-
`copy. Controls matched 2:1 to cases by age (5 years range), hospital and month of interview
`were individuals with an outpatient visit or hospitalised with a primary diagnosis that was
`neither an indication nor a known contraindication of NSAID or low dose aspirin treatment.
`The same structured questionnaire was used in all sites. Relative risk (RR) of UGIB was
`estimated using logistic regression analysis. Results: 2,777 cases and 5,532 controls have
`been included. Overall, current use of PPI (RR:0.33; 95%CI:0.27-0.39), H2-RA (RR:0.65;
`0.50-0.85) and nitrates (RR:0.52; 0.38-0.70) reduced the risk of developing an UGIB event.
`The risk reduction was stronger with PPI use among both non-aspirin NSAID (RR: 0.13;
`0.09-0.19; vs 0.30; 0.17-0.53 with H2-RA and 0.48; 0.19-1.24 with nitrates) and aspirin
`users (RR: 0.30; 0.20-0.40 vs. 0.40; 0.24-0.68 with H2-RA and 0.66; 0.44-0.98 with nitrates).
`Among individual NSAIDs, a similar risk reduction effect with PPI was observed for the 3
`most widely used (diclofenac, ibuprofen and naproxen). Among low-dose aspirin users, PPI
`(RR:0.32; 0.22-0.51) and H2-RA (RR: 0.40; 0.19-0.73)use were associated with risk reduc-
`tion, while nitrates had a weaker effect (RR: 0.69; 0.36-1.04). In patients taking clopidogrel,
`only PPI use was associated with a significant risk reduction (RR: 0.19; 0.07-0.49). However,
`among patients taking anticoagulants neither nitrates (0.67; 0.33-1.34), nor H2-RA (0.88;
`0.32-2.45) or PPI use (0.67; 0.37-1.21) were associated with a significant effect on the risk
`of UGIB event. Conclusion: Treatment with nitrates, H2-RA or PPI is associated with a
`reduction of the risk of developing UGIB events in patients taking NSAID or aspirin. However,
`only PPI therapy was associated with a marked and consistent risk reduction among patients
`receiving all types of agents including non-aspirin anti-platelet agents. Protection was much
`less apparent in patients on anticoagulant therapy.
`
`564
`
`Gastroprotection Gap: A Rising and Dangerous Omission for Elderly Users of
`NSAIDs with Arthritis
`Gurkirpal Singh, Huijian Wang, Eiichi Tanaka, Alka Mithal, Lauren Gerson, George
`Triadafilopoulos
`
`Background: GI complications from NSAIDs have been well-recognized since early 1990s,
`and were significantly reduced by the use of selective cox-2 inhibitors, concomitant proton
`pump inhibitors (PPIs) or misoprostol in the early 2000s. Recent media attention on the
`potential association of cardiovascular events with selective cox-2 inhibitors and market
`withdrawals resulted in a large decline in the use of these drugs. We report change in
`prescription patterns in a large cohort of elderly arthritis patients from January 1, 1995 to
`June 30, 2005. Methods: MediCal, the Medicaid program for California, is the largest
`Medicaid program in the US, with over 7 million participants per year. All study drugs
`were available without formulary restrictions or copayments. We studied individuals with
`physician-diagnosed arthritis who were over 65 years of age and treated with NSAIDs for
`at least 30 days. Results: Of the total 5,194,765 prescriptions for NSAIDs, 2,634,345 (50.7%)
`were for selective cox-2 inhibitors. Among the 2,560,420 prescriptions for non-selective
`NSAIDs, only 1,215,762 (47.5%) had concomitant use of PPI or misoprostol. Figure shows
`the use of selective cox-2 inhibitors and concomitant PPIs or misoprostol, as a percentage
`of all NSAID use. The increasing implementation of gastroprotection strategies over the past
`several years reached a peak in 2004 when the percent of patients not receiving gastroprotec-
`tion (Gastroprotection Gap) decreased to 14% from 91% in 1995. However, this gap more
`than doubled to 35% in 2005, following a decline in selective cox-2 inhibitor use, without
`a commensurate increase in other gastroprotective therapies. Conclusions: An increasing
`number of elderly users of NSAIDs are again left without gastroprotection. This trend, if
`left unchanged, will undoubtedly increase morbidity and mortality from NSAID-related
`complications to levels unacceptable for optimal medical care and deserves immediate
`public attention.
`
`AGAAbstracts
`
`AGA Abstracts
`
`A-82
`
`T : 89386$$CH2
`05-04-06 23:18:25
`Layout: 89386B : e
`
`Page 82
`
`Page 2 of 2
`
`Patent Owner Ex. 2017
`CFAD v. Pozen
`IPR2015-01718