Risk of bleeding with combination of aspirin and warfarin

• this Danish population based case-control study assessed the risk of serious upper gastrointestinal (GI) bleeding with antithrombotic agents, alone and in combination, and looked at trends in the use of antithrombotic drugs
  • discharge summaries of patients who were admitted to hospital between January 2000 and December 2004 with a main diagnosis of peptic ulcer or gastritis were reviewed
    • cases were included in the study if they had significant bleeding (melaena, subnormal haemoglobin, transfusion) and a potential bleeding source in the stomach or duodenum identified on endoscopy or surgery
    • investigators were blinded to the exposure of the subjects to antithrombotic agents
    • 1443 cases identified were age- and sex-matched to 57,720 controls who had not been admitted for GI bleeding
    • main outcome measure was exposure to low-dose aspirin, clopidogrel, dipyridamole, vitamin K antagonists (e.g. warfarin) and combined antithrombotic treatment. Adjustments were made for potential confounders including current drug use (e.g. non-steroidal anti-inflammatory drugs [NSAIDs], selective serotonin reuptake inhibitors [SSRIs], anti-ulcer drugs), and past diagnoses (e.g. peptic ulcer, upper GI bleeding, diabetes, ischaemic heart disease)
  • results
    • adjusted odds ratios associating drug use with upper gastrointestinal bleeding were
      • 1.8 (95% confidence interval 1.5 to 2.1) for low dose aspirin
      • 1.1 (0.6 to 2.1) for clopidogrel
      • 1.9 (1.3 to 2.8) for dipyridamole
      • 1.8 (1.3 to 2.4) for vitamin K antagonists
      • corresponding figures for combined use were
        • 7.4 (3.5 to 15) for clopidogrel and aspirin
        • 5.3 (2.9 to 9.5) for vitamin K antagonists and aspirin,
        • 2.3 (1.7 to 3.3) for dipyridamole and aspirin
        • number of treatment years needed to produce one excess case varied from 124 for the clopidogrel-aspirin combination to 8800 for clopidogrel alone

Conclusions:

• combinations of aspirin plus clopidogrel and aspirin plus warfarin are associated with particularly high rates of serious upper gastrointestinal bleeding
• these combinations should be used with care, as the risk of major bleeding may outweigh any benefits (e.g. reductions in cardiovascular events)
• this case-control study (1) suggests that combinations of aspirin plus clopidogrel and aspirin plus warfarin should only be used with caution for conditions where it has been established that the benefits outweigh the increased risk of bleeding, and then only for a limited time. The risk of bleeding with aspirin plus dipyridamole appears lower.

Notes:

• main limitation of this study is its design
  • retrospective case-control studies cannot demonstrate causality. In this study, antithrombotic drugs are associated with serious upper GI bleeds, but this does not mean that they are definitely the cause of those bleeds.
  • misallocation of subjects when investigators are considering which cases to include is a potential source of bias
  • another major limitation of the study is that there were only a small number of people in many of the sub-groups
  • data were not available for all potential confounders (e.g. smoking, high alcohol intake, use of over-the-counter NSAIDs)
  • selection bias may also have been possible as patients who were taking antithrombotic therapy may have been more likely to be admitted when they had symptoms of a GI bleed than those who were not

Reference:

1. Hallas J et al. Use of single and combined antithrombotic therapy and risk of serious upper gastrointestinal bleeding: population based case-control study. BMJ. 2006 Oct 7;333(7571):726