Incidence and possible risk factors for clinical upper gastrointestinal events in patients taking selective cyclooxygenase-2 inhibitors: A prospective, observational, cohort study in Taiwan.

Background: Use of selective cyclooxygenase (COX)-2 inhibitors is associated with a better gastrointestinal (GI) safety profile than use of other NSAIDs. However, the risk factors for clinical upper GI events (symptomatic ulcers and ulcer complications) in COX-2 inhibitor users have been rarely evaluated in the past.

Objective: The aim of this study was to assess the annual incidence of and identify the risk factors for clinical upper GI events in chronic COX-2 inhibitor (celecoxib and etoricoxib) users.

Methods: A prospective, hospital-based, observational cohort study was conducted in patients taking COX-2 inhibitors (celecoxib or etoricoxib) without comorbidity. COX-2 inhibitor prescription was conducted according to the guidelines issued by the Taiwan National Health Insurance and was identified by the computerized prescribing system of the Taipei Veterans General Hospital, Taipei, Taiwan. Patients with cardiovascular disease, pulmonary, hepatic or renal insufficiency, or malignancy were excluded. Patients received regular follow-up once a month. Between visits, patients were asked to report to the outpatient clinic if they had persistent ulcer or GI symptoms (dyspepsia, vomiting, nausea, heartburn, or acid regurgitation) not relieved by antacids for 1 week, or to the emergency department if they had evidence of GI bleeding or ulcer complications (melena, hematemesis, hematochezia, or sudden onset of severe epigastric pain). Endoscopy was performed to document any gastroduodenal ulcers with or without ulcer complications. The primary end point was the annual incidence and the significant risk factors for clinical upper GI events (symptomatic ulcers and ulcer complications).

Results: A total of 1158 COX-2 inhibitor users were identified; 96 refused to participate and 129 were excluded. The mean (SD) age of the remaining 933 COX-2 inhibitor users was 69 (15) years with 528 women (56.6%) and 405 men (43.4%). Mean time of follow-up was 12.4 months. The annual incidence of clinical upper GI events in these patients taking COX-2 inhibitors was 4.6% (44 events/959 patient-years), with symptomatic ulcers in 3.6% and ulcer complications in 1.0%. Multivariate logistic regression analysis found that a history of peptic ulcer disease (PUD) (odds ratio [OR = 4.61; 95% CI, 1.86-11.40; P = 0.001), concomitant use of steroids (OR = 2.99; 95% CI, 1.39-6.46; P = 0.005), aspirin (OR = 13.47; 95% CI, 5.89-30.82; P < 0.001), and other NSAIDs (OR = 60.49; 95% CI, 11.93-306.64; P < 0.001) were significant independent risk factors for clinical upper GI events in these patients taking COX-2 inhibitors. Age >60 years was not found to be a risk factor.

Conclusions: The annual incidence of clinical upper GI events was 4.6% in these Taiwanese patients without comorbidity taking COX-2 inhibitors. A history of PUD and concomitant use of steroids, aspirin, or other NSAIDs, were found to be significant risk factors for clinical upper GI events in these patients.