Independent predictors of poor vitamin K antagonist control in venous thromboembolism patients. Data from the EINSTEIN-DVT and PE studies.

Vitamin K antagonists (VKA) are used to prevent recurrent disease in patients with venous thromboembolism (VTE). Their efficacy and safety depend on individual time in therapeutic range (iTTR) and variability of International Normalised Ratios (INR). We aimed to identify independent predictors of poor VKA control > 28 days. In a prospective cohort of 3825 VTE patients, separate logistic regression analyses were performed to identify predictors of low iTTR (first quartile) and instability (iTTR median). Subsequently, the association between these predictors and clinical outcomes was investigated. Weight < 50 kg (odds ratio [OR]=1.89; 95 % confidence interval [CI] 1.03-3.49), active cancer at baseline (OR=1.52; CI1.05-2.19), secondary VTE (OR=1.42; CI1.20-1.68), and INR < 2.0 at stop of double therapy (OR=1.35; CI1.09-1.67) were independent predictors of low iTTR. The first two were also predictive for instability (OR=1.96; CI1.06-3.63 and OR=1.95; CI1.36-2.80, respectively). ORs of early (≤ 28 days) low iTTR and instability depended on VKA type. In acenocoumarol users, early low iTTR was an independent predictor of subsequent low iTTR (OR=1.92; CI1.31-2.80) and instability (OR=1.55; CI1.07-2.23). In warfarin users, early low iTTR (OR=1.36; CI1.09-1.69) and instability (OR=1.25; CI1.01-1.55) were additionally predictive for low iTTR, but only the latter was predictive for instability (OR=1.91; CI1.57-2.32). Many predictors of VKA control also predicted premature discontinuation, but only region was prognostic for clinical outcome. In conclusion, we identified several independent predictors of low iTTR and instability > 28 days, which showed some similarities but did not fully overlap. Early VKA control was of additional value for prediction of both, but had to be interpreted in the context of VKA type.