1. Recognize the risk factors for GI complications stemming from use of NSAIDs.

2. Adopt the major strategies for reducing the risk for GI complications related to use of NSAIDs.

3. Discuss the results from major trials of cyclooxygenase (COX)-2 inhibitors.

4. Predict future changes in the gastroenterology work force.

5. Discuss possible colonoscopy alternatives and new colonoscopy designs.

NSAID-associated GI Disorders and Management
Jay L. Goldstein, MD, Professor of Medicine and Vice Head for Clinical Affairs, Department of Medicine, University of Illinois at Chicago College of Medicine; Staff Physician, University of Illinois Hospital and Jesse Brown VA Medical Center, Chicago

Gastrointestinal (GI) side effects associated with nonsteroidal anti-inflammatory drugs (NSAIDs): range from mild dyspepsia to fatal GI complications; case—woman, 51 yr of age, with right hip and left knee pain, asymmetric; has large-joint osteoarthritis; limited in her activities to walking 1 block on level ground; treated with naproxen 500 mg bid for past 3 yr after failing to have symptoms controlled with 4 g of acetaminophen daily; no drug-related problems noted to date; patient feels better; has negative family history and no drug allergies; physical examination normal; negative fecal occult blood test (Hemoccult); patient doing well and wants to continue NSAID therapy; discussion—average- risk person taking nonselective NSAID for 1 yr runs risk of developing ulcer complication (primarily bleeding; 80%) or symptomatic ulcer; 1% to 2% develop complications, and 1% to 2% develop symptomatic ulcers; in patient who develops ulcer or ulcer complication, prospective randomized trials show that 8 wk of therapy with proton pump inhibitor (PPI) superior to histamine (H)2 blocker; H2 blockers have no role in treatment or prevention of NSAID-related ulcers

Preventing GI complications from NSAID use: goals include identifying those at high risk and taking appropriate action; risk factors—history of upper GI clinical event (increases risk 4-fold); older age (>60 yr of age); anticoagulants; corticosteroid therapy (combined with NSAID increases risk 2-fold); selective serotonin reuptake inhibitors (2-fold increased risk); high-dose or multiple NSAIDs (3 groups include those who take NSAID more frequently than prescribed, those who combine over-the-counter NSAID with prescribed NSAID, and aspirin users); misoprostol ulcer complications outcomes and safety assessment (MUCOSA) study showed that as number of risk factors increases, growth in relative risk (RR) almost exponential; aspirin alone causes ulcers and ulcer complications (compared to general population not using it), but risk even greater if combined with NSAID; misconception that adaptation to NSAID toxicity seen (no plateau seen; 2 studies showed that RR constant); major strategies for reducing risk include using gastroprotective agents (eg, misoprostol, PPI) or safer NSAIDs (eg, cyclooxygenase [COX]-2 inhibitors)

Efficacy of various preventive strategies: in prospective randomized controlled trials, shown that misoprostol in NSAID users reduces rate of endoscopic ulcers in stomach and duodenum and of ulcer complications; H2 blockers ineffective; PPIs shown to reduce gastric and duodenal ulcers significantly and likely to reduce complications; misoprostol—side effect diarrhea (change in bowel habits); causes ≈40% reduction in events; effective in those able to tolerate it; PPIs—more commonly used; data show that whether pantoprazole 20 or 40 mg or omeprazole 20 mg used, majority of patients on NSAIDs remained ulcer-free over time; placebo-controlled data showed significant reduction of endoscopic ulcers with coprescription of PPI; Hong Kong study—patients on aspirin with bleeding, treated with PPI, and had Helicobacter pylori eradicated; patients endoscoped and proven to have healed ulcers; mean age 65 yr; resumed on low-dose aspirin alone or with lansoprazole; PPI reduced rate of recurrent upper GI ulcer complications (rebleeding) over 12 mo

Adherence to protective strategies during NSAID therapy: Abraham study—found that in patients with one risk factor, only 27% received COX-2 inhibitor, PPI, or misoprostol; even with 3 risk factors, <50% of patients recognized to have problem; study—found that for every 4 days of NSAID use, patients taking equivalent of only 3 days of PPI; more than one-third of patients not adherent to coprescribed PPI; study—found that approximately one-third of patients taking <80% of coprescribed PPI; as adherence declines from 100% to 0%, rate of events increases

COX-2 inhibitors: arachidonic acid (AA) precursor to prostaglandin; 2 COX enzymes (COX-1 and COX-2); COX-1— present in all tissues and converts AA into prostaglandins in GI mucosa; prostaglandins protect GI mucosa by increasing bicarbonate secretion, accelerating cellular repair, and maintaining mucus secretion; in platelets, AA converted into thromboxane, causing hemostasis; agent that inhibits COX-1 (eg, aspirin) leads to loss of GI protection (leaving it vulnerable to injury) and antiplatelet effect; COX-2—expressed in sites of inflammation; initiates, maintains, and potentiates inflammatory response (therapeutic target of NSAIDs); use of nonspecific NSAID that inhibits COX-1 and COX-2 equally results in anti-inflammatory effect; use of COX-2—specific agent at therapeutic dose does not inhibit COX-1 (no antiplatelet effect) but inhibits COX-2, resulting in anti-inflammatory response; when inhibiting COX-2 by using traditional NSAID or COX-2 inhibitor, degree of inhibition and therapeutic response same; agents include—lumiracoxib (not approved by Food and Drug Administration [FDA]), etoricoxib (not approved by FDA), rofecoxib (Vioxx; withdrawn), valdecoxib (withdrawn), and celecoxib (only one remaining); Vioxx Gastrointestinal Outcomes Research (VIGOR) trial—compared patients on naproxen 500 mg bid to rofecoxib 50 mg and found rate of events lower with rofecoxib (≈50% reduction in rate of ulcer complications); when stratifying risk, RR reduction similar, but absolute risk so much greater in high-risk population that number needed to treat (NNT) decreases, as risk in population increases (basis for cost-effectiveness argument to reserve coxibs for high-risk populations); coxibs work in low-risk patients, but NNT higher; Celecoxib Long-term Arthritis Safety Study (CLASS)—also found ≈50% reduction in rate of symptomatic ulcers and ulcer complications with coxib in nonaspirin users; >50% reduction in rate of ulcer complications with lumiracoxib in nonaspirin users; in aspirin-using population, however, benefit of coxib lost and performance no different from that of NSAID

Upper GI bleeding risk related to aspirin prophylaxis: case continued—patient returns 1 yr later; has hypertension and new-onset type 2 diabetes (diet-controlled); placed on β-blocker and 81 mg of aspirin daily; osteoarthritis stable; clear indication for prophylaxis present because of increased risk (2- to 4-fold) for upper GI bleeding with aspirin; study data—in Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET) and CLASS trial, patients on aspirin lost benefit of coxib; one endoscopic trial suggested some benefit to using aspirin with coxib (rather than naproxen with coxib; more trials necessary); PPIs alone reduce endoscopic ulcer rates for those on aspirin only; in speaker’s study comparing patients on naproxen, aspirin, and PPI with patients on coxib, aspirin, and PPI, endoscopic ulcer rates same; patients on clopidogrel did worse than patients on aspirin and PPI; case continued—patient did well for 5 yr on naproxen and aspirin, but then developed upper GI bleeding; hospitalized and required 3 U of packed red blood cells (RBCs); upper endoscopy revealed 1-cm deep bleeding gastric ulcer with visible vessel; no rebleeding observed after endoscopic therapy; patient treated with NSAID and PPI and continued on aspirin; endoscopy 8 wk later revealed healed ulcer; study by Chan—study of patients who had bleeding with NSAID, had H pylori eradicated, and treated with PPI; healing documented (normal endoscopy); patients randomized to receive celecoxib alone or diclofenac and omeprazole; followed for 6 mo; found no difference between 2 approaches; however, rate of recurrent ulcer bleeding 5%; with COX-1 inhibitors, patients twice as likely to bleed from colon, with increased risk of bleeding from small bowel (accounts for many of esophagogastroduodenoscopy [EGD]-negative patients who present with iron-deficiency anemia and have undergone appropriate gynecologic examinations); capsule endoscopy study—in healthy volunteers, incidence of mucosal breaks higher with naproxen (≈9 fold) than with celecoxib, proving that coxibs safer in small bowel

FitzGerald hypothesis: 2 opposing forces (COX-1 and COX-2); platelet COX-1 causes platelet aggregation (prothrombogenic); in endothelial cells (COX-2), results in generation of prostacyclins, which cause vasodilatation and have antithrombotic effect; under normal conditions, opposing forces balance out; thought that if using NSAID that inhibits both pathways, balance preserved and no net cardiovascular (CV) risk; if COX-2 inhibitor used, protective (prostacyclin) pathway removed, leaving platelets intact (net prothrombogenic state); CLASS trial stated that rate of CV events same whether on NSAID or coxib; in VIGOR trial, patient on rofecoxib 2.4 times more likely to have thrombotic event than patient on naproxen (no placebo arm); Adenomatous Polyp Prevention on Vioxx (APPROVe) trial—looked at effect of long- term COX-2 inhibition in patients at high risk for recurrence of adenomatous polyp; placebo-controlled; found patients twice as likely to develop cardiac event; led to removal of rofecoxib from market; study—compared use of celecoxib 400 mg bid to celecoxib 200 mg bid to placebo; dose response not significant, but response to celecoxib 400 mg significantly higher than to placebo; whether patient aspirin user or not, hazard ratio increased (although not statistically significant), suggesting protective effect of aspirin against CV risk associated with celecoxib; in meta-analyses, varying toxicities with different COX-2 inhibitors seen (more studies necessary); consensus that naproxen not cardiotoxic; concept of COX-1 selectivity or nonselectivity as determinant of CV toxicity too simplistic; all COX inhibitors have some heightened risk, and each one requires evaluation in prospective trials; Multinational Etoricoxib and Diclofenac Arthritis Long-term (MEDAL) trial—concluded that COX-2 inhibitors do not have heightened CV risk

Megatrends in GI: Where We Are and Where We Are Likely to Go
Grace H. Elta, MD, Professor, Department of Internal Medicine, Associate Chief for Clinical Programs, Director, Medical Procedures Unit, and Director, Pancreatic and Biliary Disorders, University of Michigan Medical School, Ann Arbor

Endoscopy in United States: in 2007, ≈20 million endoscopies performed annually, of which 70% colonoscopies; 3- to 4-fold increase since 1989; site of service—office (9%), ambulatory endoscopy center or ambulatory surgery center (36%), and hospital (55%); 35% to 40% of endoscopies use anesthesia providers; physicians performing—68% by gastroenterologists; 20% by general surgeons or colorectal surgeons; 12% by primary care physicians (data from 2003 Medicare claims database)

Gastroenterology work force: ≈12,000 valid American Board of Internal Medicine (ABIM) certificates; ≈5% of gastroenterologists never pass board examination (noncertified but practicing; ≈12,000 gastroenterologists in United States (does not include non-GI endoscopists); number of gastroenterologists expected to remain flat (no funding for more GI fellowship positions); predictions—number of colonoscopies will increase, as more eligible patients screened; number of gastroenterologists will increase slightly (at slower pace than procedures); outcome determined by reimbursement (difficult to predict); if reimbursements same or decrease slightly, more time devoted to procedures by gastroenterologists; if reimbursements decrease, may see procedures performed by technicians or midlevel providers under supervision of physician; also possible that alternative imaging techniques (eg, computed tomography [CT] colonography) will take some of work load

Diagnostic colonoscopy: screening and polyp surveillance comprises 50% to 60% of colonoscopies in patients >50 yr of age; question of whether new diagnostic or surveillance tests will replace some of volume; number of eligible patients screened depends on location (should reach levels of mammography [80%]); question of whether simpler techniques increase percentage of people screened; colonoscopy alternatives—stool, serum genetic, and proteomic assays; currently lack accuracy; possibly important in 10 to 20 yr

CT colonography (CTC): large variation in polyp sensitivity; category 1 code possible in future; current problems— still requires full preparation; radiation dose significant (equivalent to 350 chest x-rays); interpretation time-consuming; due to cost issues, American College of Radiology stated that polyps 6 to 9 mm in size (ie, low-risk polyps) should not be sent for colonoscopy; however, 6% of polyps in that size range have advanced histology; presence of extracolonic findings; virtual colonoscopy (VC)—in Madison, Wisconsin, 50% of primary care physicians and patients choosing VC; detects same number of advanced adenomas (≥1 cm) as standard colonoscopy; standard colonoscopy finds more medium- and small-sized polyps than CTC

Colon capsule: disallowed for clearance by FDA; data from Europe indicated positive capsule study if polyp ≥6 mm or ≥3 polyps of any size; 15% to 26% of capsules failed to reach anus; capsule found 70% to 77% of colonoscopy findings; requires extensive preparation; has false-positive findings

New colonoscope design: Aer-O-Scope—computer-adjusted balloon; self-navigates around colon; in preliminary trial, reached cecum in 10 of 12 patients; images can be viewed at later time; Invendoscope—has inverted sleeve technology that creates fewer loops; in pilot study with 28 patients, cecal intubation rate poor and some patients experienced pain

Diagnostic colonoscopy replacement: American Gastroenterological Association (AGA) favors CTC; colon capsule thought not accurate enough; stool and serum studies ideal but not in near future; nothing close to market in new colonoscope design

Endoscopic sedation: good patient care results when anesthesia providers give propofol; question of whether added benefit to patient care worth cost; fear of reevaluation of base codes if anesthesia providers continue; payors starting to deny payment; gastroenterologist and registered nurse-delivered propofol—rarely performed, due to restricted label and state nursing boards, but just as safe as other sedation methods; fospropofol (Aquavan)—prodrug for propofol; long half-life; possibly riskier than propofol; seeking FDA approval without restricted label; computer-assisted personal sedation (CAPS)—awaiting approval from FDA; fairly complex system; cost problematic with all devices

New technologies for gastroenterologists: virtual histology—possibly beneficial in Barrett’s esophagus and inflammatory bowel disease; possibly higher polyp detection rates; third eye endoscope—has increased polyp detection but widescale use uncertain, due to cost and inconvenience; natural orifice transluminal endoscopic surgery (NOTES)— may bring new tools; balloon enteroscopy—revolutionized small bowel endoscopy

Predictions: diagnostic colonoscopy likely replaced at some point; increasing complexity of endoscopy (new image- enhancement tools; NOTES); plentiful work for gastroenterologists (Medicare expected to double by 2045, with only slight increase in gastroenterology work force)