Prevalence of drug-drug interactions and clinical relevance for treatment with oral anticoagulants

Abstract

The aim of this thesis was to describe prevalences and frequencies of potential DDIs in the Swedish outpatient population, and the clinical effects in treatment with oral anticoagulants.

Two cross-sectional studies were conducted to establish the prevalence and frequency of potential DDIs among Swedish outpatients. Study I included the whole outpatient population in Sweden, and study II the pediatric outpatient population. The prevalence of clinically relevant potential DDIs among patients with at least two drugs was 19 % and 1.4 % respectively. The proportions of clinically relevant DDIs that potentially lead to reduced treatment effect were 49 % of class D interactions (recommendation to avoid) and 54 % of class C interactions (may require e.g. dose adjustment) in the whole population. The corresponding proportions were and 48 % and 32 % in the pediatric population. A limited number of drugs were involved in a large proportion of potential DDIs. Furthermore, many of the clinically relevant DDIs may lead to reduced treatment effect, an aspect of interactions that may be underestimated in clinical practice.

In study III, a cohort of warfarin patients was studied to analyze the longitudinal effects of initiation of amiodarone therapy on warfarin dose and INR. The mean weekly warfarin dose was 24.6 % lower after initiation of amiodarone (95 % CI, 23.5–25.6 %; P < 0.001). Mean weekly INR peaked the third week of concomitant treatment. The fraction of patients with an INR over 3 was 37.1 % at that point, compared to 11.7 % at baseline.

The increased risk of bleeding or thromboembolism, potentially associated with DDIs among patients treated with non-vitamin K antagonist oral anticoagulants (NOACs), was investigated with survival analyses in study IV. Atrial fibrillation outpatients were included in the cohort. Compared to patients not exposed to the interacting group of drugs, exposure to potential pharmacodynamic DDIs were associated with an increased risk of any severe bleed, for patients with apixaban HR (95 % CI) 1.47 (1.33-1.63), rivaroxaban 1.7 (1.49-1.92), and dabigatran 1.26 (1.05-1.52). In addition, exposure to CYP3A4 and/or P-gp inhibitors was associated with an increased risk of any severe bleed for patients treated with apixaban 1.23 (1.01-1.5). No significant effects could be established for patients exposed to inducers of CYP3A4 and/or P-gp.

In conclusion, one fifth of patients with at least two drugs in the whole Swedish outpatient population was exposed to potential DDIs. The identified drugs and potential clinical consequences need to be considered in clinical practice to avoid adverse events. Patients initiated on amiodarone during warfarin treatment had a mean dose reduction of 25 % and close monitoring during the first weeks is important to avoid supratherapeutic anticoagulant effect. Pharmacodynamic and pharmacokinetic DDIs with NOACs expose atrial fibrillation patients to increased risk of bleeding. These DDIs are important to consider from a risk benefit assessment perspective in patients treated with NOACs.