Patients
Patients admitted to the Bejing Anzhen Hospital of Capital Medical University from July 27, 2018, to July 1, 2019, who underwent aortic surgery, were enrolled in this study. Consecutive patients were assessed for eligibility for enrollment based on the following inclusion criteria: patients who had indications for warfarin and underwent aortic surgery. The exclusion criteria were as follows: active bleeding or bleeding diseases, hematologic disorder, severe hepatic or renal insufficiency (creatinine clearance < 30 mL/min), and unable or unwilling to give written informed consent. The study had been approved by the Medical Clinical Research Ethics Committee of Bejing Anzhen Hospital.
Data Collection
Anticoagulation stabilization is defined as the warfarin dose being unchanged for 14 days and the INR value is stable within the range of 80-120% of the INR target value for at least 7 days[21,22]. The target range of INR after mechanical aortic valve replacement is 1.8-2.2. The target range of INR after mechanical mitral valve replacement or double valve replacement is 2.0-2.5. Stabilization INR (INRss) is defined as the second INR value after anticoagulation stabilization[21,22]. The observed warfarin maintenance dose was defined as the dose of warfarin when the patient achieved INRss.
Patient demographics (age, gender, height and body weight), smoking and drinking history, bleeding history and thrombosis history, comorbidities, type of surgery, concomitant medications, gene polymorphism (CYP2C9 and VKORC1), baseline INR observation, target INR range, the full warfarin dosing history and all available INR observations were recorded from the first day of warfarin therapy until a stable observed INR (INRss).
Initial warfarin dose and subsequent warfarin dose adjustment were decided by clinicians who were not involved in the study. The initial dose of warfarin was usually 3mg (one tablet, Orion Corporation Orion Pharma) according to the routine dose in the Chinese population. 42 (95.5%) patients in this study received an initial warfarin dose of 3mg. And two other patients were treated with an initial warfarin dose of 1.5mg due to the high initial INR values. Dose adjustment of warfarin was based on the experience of clinicians instead of any warfarin dosing prediction algorithms.
Statistical Analysis
The predicted warfarin doses of the IWPC and Gage algorithms were calculated by the website of http://www.WarfarinDosing.org[23]. The predicted warfarin dose of the Bayesian-priori (dose estimation before starting therapy) and Bayesian-posteriori (dose estimation after initiating therapy) algorithm were calculated by the developed freely available software Bayesian dosing tool (Warfarin Dose Calculator) which was written in Java and had been published[18].
The predictive performance of each dosing method was compared by calculating mean prediction error (MPE), mean squared error (MSE), root-mean-squared error (RMSE), and the percentage of patients whose predicted warfarin dose fell within 20% of the maintenance dose [24,25]. MPE, MSE and RMSE were calculated as follows[22]:
N is the number of patients. PEi is the predicted minus observed maintenance dose (prediction error) of the i-th individual. MSE is the average of the sum of squared differences between predicted minus observed maintenance dose. RMSE is the square root of MSE. MPE is an estimate of the bias, MSE and RMSE are an estimate of the imprecision (milligrams per day).
In the subgroup analysis, patients with a maintenance dose of warfarin between 2.25mg and 3.75mg were divided into conventional dose group, and the others were divided into the unconventional dose group. The predictive performance of each dosing method in each subgroup was compared. Pearson’s correlation analysis was conducted to compare the relationship between the predicted and the actual warfarin dose.
Definition of bleeding and thrombotic events
Thrombotic events included valve thrombosis, coronary embolism, peripheral embolism, pulmonary embolism, ischemic stroke and transient ischemic attack. Valve thrombosis was defined as impairment of the valve due to thrombus deposition, recorded by Doppler ultrasound or during surgery. Coronary, peripheral and pulmonary embolic events were indicated by Doppler ultrasound, angiography, or surgery. Cerebral thromboembolic events included the onset of a transient or definitive symptomatic nervous system stroke and/or computed tomography brain scan showing evidence of an ischemic cerebrovascular sequela[26].
Major bleeding events were defined according to the criteria of International Society of Thrombosis and Haemostasis (ISTH) as clinically overt bleeding which was fatal or associated with any of the following: i) a fall in hemoglobin level of 20 g/l or more or documented transfusion of at least 2 units of whole blood or red cells, ii) involvement of a critical area or organ (intracranial, spinal, ocular, pericardial, articular, intramuscular with compartment syndrome, retroperitoneal)[27]. All other bleeding events were considered minor.