Anonymised patient data from the records of 1,500 GP practices using EMIS Web has enabled a major new study for the National Institute for Health Research (NIHR) into the risks and benefits of modern anticoagulant drugs.
Researchers used the QResearch database – which draws on routinely captured data from participating EMIS practices – for their study comparing the health outcomes of patients treated with three commonly-prescribed direct oral anticoagulants (DOACs) with those treated with warfarin.
Published in the BMJ, the study gives the most comprehensive overall picture of the real-world use of anticoagulants in patients with a range of conditions, not just atrial fibrillation. Researchers analysed data from the primary care records of almost 200,000 people, comparing two large groups over a five-year period. One group had abnormal heart rhythm, the other a range of other conditions treated with anticoagulants.
The results of the study will provide important new evidence on the safety profile of the different anticoagulant medications and help doctors make better prescribing decisions.
The study found:
- apixaban was the safest drug, with reduced risks of major bleeds, intracranial and gastro-intestinal bleeding compared with warfarin
- in the group with abnormal heart rhythm, dabigatran was linked to a reduced risk of intracranial bleed, while rivaroxaban was linked to a reduced risk in the group with other conditions
- rivaroxaban and low dose apixaban were associated with increased risks of all-cause mortality compared with warfarin in both groups of patients.
More work is needed to understand the implications of some of the findings for real world clinical practice. For example, DOACs are not suitable for all patients who require anticoagulants - prosthetic valve cases still require warfarin.
QResearch® is a non-profit making venture run by the University of Nottingham in collaboration with EMIS Health. More than 1,500 EMIS practices, representing around 30 million patient lives, contribute anonymised data to the database.