Underutilization of Gastroprotective Strategies in Aspirin Users at Increased Risk of Upper Gastrointestinal Complications

L. E. Targownik; C. J. Metge; S. Leung


Aliment Pharmacol Ther. 2008;28(1):88-96. 

In This Article


In Manitoba, the provincial Ministry for Health (Manitoba Health) maintains comprehensive healthcare utilization databases on all residents of Manitoba. These contain information on resident demographics, all resident hospital and physician encounters since 1984, and all outpatient prescriptions dispensed since 1995.

Before March 2004, the 9th edition of the International Classification of Disease (ICD-9) disease definitions were used to codify diagnoses, whereas ICD-10 is used for codification from April 2004 onwards. All inpatient hospitalizations are coded using up to 16 full five-character ICD-9 or ICD-10 codes, whereas physician encounters are associated with one abbreviated three-character ICD-9 diagnosis per visit.

We identified a cohort of subjects with likely CV disease by searching for all subjects with a hospital admission between 1 April 2002 and 31 March 2006 coded for any of: acute or previous myocardial infarction, unstable angina, nonhaemorrhagic cerebrovascular accident, coronary artery bypass grafting, coronary angioplasty/coronary stenting or carotid endarterectomy. This cohort was stratified into two groups based on whether or not they have had a previous admission to hospital with a diagnosis consistent with PUD and/or one of its complications.

We then distributed a survey package to a random sample of 3500 patients meeting the diagnostic criteria for CV disease and no PUD, and to all identified subjects with both CV disease and a prior history of hospitalization for PUD, to ensure that we had adequate representation of subjects at high risk of GI complications. The main purpose of the survey was to gather data concerning the use of over-the-counter medications, specifically aspirin, ibuprofen and antacids and to collect information on each subject's indications for aspirin use, gastrointestinal symptom burden, and understanding of the side effects of aspirin therapy. The survey was created with the assistance of an expert in survey design with the intent of it being understood by a subject at a sixth-grade reading level, and which would take no more than 10 min to complete. A copy of the survey can be seen in Appendix S1.

Distribution of the survey packages was performed through an independent third-party mail-out service so as to blind investigators from the identity of potential subjects. The survey package was composed of the survey letter, a cover letter explaining the purpose of the study, a copy of a consent form to be signed and a preaddressed, postage-paid envelope for return of the survey and a signed consent form. Potential recruits were also provided with a pen and a $2 cash incentive. The initial mail-out took place in August 2006, and a second mail-out took place for nonresponders in October 2006. The survey contained an entry field in which the respondent could enter the date of survey completion.

Data collected from the returned surveys were linked to demographic and healthcare utilization information in the previously described provincial healthcare databases. We specifically tracked the use of certain medications, including prescription NSAIDs and COX-2 inhibitors, histamine-2-receptor antagonists, PPIs, warfarin, clopidogrel, and systemic corticosteroids. A subject was considered to be an active user of the medication if there was a dispensation within 90 days before the date the survey was completed. We also identified data concerning medical comorbidities, including respiratory, autoimmune, hepatic, renal, or active malignancy (excluding nonmelanoma skin cancers). A subject was considered to have a comorbid condition if there were at least two outpatient healthcare encounters or one inpatient healthcare encounter from that category in the year before the date of completion of the survey.

First, we compared the prevalence of self-reported aspirin use among subjects with cardiac disease between those with a history of hospital admission for PUD and those with no similar history. We also compared the prevalence of PPI use rates between subjects with and subjects without other risk factors for aspirin-related UGI complications, adjusting for the differential sampling of PUD and non-PUD subjects. The risk factors of interest were significant active medical comorbidity, use of nonaspirin NSAIDs, systemic corticosteroids, warfarin, clopidogrel, or over the age of 70. We used multivariate logistic regression to determine if any of these variables was predictive of PPI utilization in aspirin users. Chi-square testing or the Student's t-test was used to assess the significance of comparisons between categorical or continuous variables, with P < 0.05 being considered statistically significant.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.