Background: Patients with indications for anticoagulation are increasingly treated with combinations of antiplatelet medications and vitamin K antagonists, but the safety of such combinations is unknown. Hallas and colleagues used a large pharmacologic database and hospital data system to study serious gastrointestinal bleeding complications of aggressive antithrombotic therapy.
The Study: The databases covered all 470,000 residents of one county in Denmark. Cases of serious gastrointestinal bleeding were defined as hospital admission for significant melena, subnormal hemoglobin, or need for transfusion plus a gastrointestinal source of bleeding identified by surgery or endoscopy. The pharmacy records of the 1,443 validated patients and those of more than 57,000 control patients were examined for exposure to aspirin, clopidogrel (Plavix), dipyridamole (Persantine), and vitamin K antagonists. Pharmacy records also were examined for use of nonsteroidal anti-inflammatory drugs (NSAIDs), selective serotonin reuptake inhibitors, systemic corticosteroids, antiulcer medications, nitrate vasodilators, or disulfiram (Anta-buse). Medical records were examined for history of upper gastrointestinal bleeding, peptic ulcer disease, diabetes, ischemic heart disease, or any alcohol-related condition. All 830,000 prescriptions for the population between 1995 and 2004 were reviewed.
Results: Patients with gastrointestinal bleeding had more chronic diseases and greater use of gastrointestinal irritants such as NSAIDs and other medications than control patients. Overall, 380 (26.3 percent) of patients were exposed to at least one of the four groups of antithrombotic medications. The odds ratios for current exposures to antithrombotic medication were all statistically significant. After adjustment for chronic diseases and use of medications related to gastrointestinal bleeding, the odds ratio for clopidogrel did not reach statistical significance. The adjusted odds ratios for previous use of anti-thrombotic medications were not statistically significant except for history of vitamin K antagonist use. The association with gastrointestinal bleeding was much stronger when combinations of antithrombotic medications were being used. The combined effect was much greater than the sum of the two individual medications (see accompanying table) and was particularly large for the combination of aspirin and clopidogrel. The authors estimated the number of treatment-years needed to produce one additional bleeding episode was 124 for clopidogrel/aspirin, compared with 1,040 for vitamin K antagonist monotherapy and 8,800 for clopidogrel monotherapy. The exposure to combined regimens rose from 758 person-years in 2000 to 3,978 in 2004.
|Medication||Adjusted odds ratio||Crude odds ratio|
|Clopidogrel (Plavix) alone||1.1||3.1|
|Vitamin K antagonist alone||1.8||2.2|
|Dipyridamole (Persantine) alone||1.9||2.4|
|Aspirin plus clopidogrel||7.4||12.6|
|Aspirin plus vitamin K antagonist||5.3||6.4|
|Aspirin plus dipyridamole||2.3||3.1|
Conclusion: The authors conclude that combination antithrombotic therapy is increasingly common and is associated with high rates of gastrointestinal bleeding.
editor's note: This study is particularly useful to family physicians because it was conducted in the “real world” of community practice without the extensive patient exclusions that are common in clinical trials.1,2 In addition to being more typical of the patients with multiple comorbidities seen in family practice, patients in the study had access to over-the-counter medications. This makes the crude odds ratio for aspirin plus clopidogrel particularly frightening. As we aim for better control of international normalized ratios, physicians must remain vigilant in watching for adverse effects, particularly subtle indications of gastrointestinal bleeding.—a.d.w.