Because of the narrow therapeutic window and huge inter-individual variation, the individual precision on anticoagulant therapy of warfarin is challenging. In our study, we aimed to construct a Back Propagation Neural Network (BPNN) model to predict the individual warfarin maintenance dose among Chinese patients who have undergone heart valve replacement, and validate its prediction accuracy. In this study, we analyzed 13,639 eligible patients extracted from the Chinese Low Intensity Anticoagulant Therapy after Heart Valve Replacement database, which collected data on patients using warfarin after heart valve replacement from 15 centers all over China. Ten percent of patients who were finally enrolled in the database were used as the external validation, while the remaining were randomly divided into the training and internal validation groups at a ratio of 3:1. Input variables were selected by univariate analysis of the general linear model; 2.0, the mean value of the international normalized ratio (INR) range 1.5–2.5, was used as the mandatory variable. The BPNN model and the multiple linear regression (MLR) model were constructed by the training group and validated through comparisons of the mean absolute error (MAE), mean squared error (MSE), root mean squared error (RMSE), and ideal predicted percentage. Finally, 10 input variables were selected and a three-layer BPNN model was constructed. In the BPNN model, the value of MAE (0.688 mg/day and 0.740 mg/day in internal and external validation, respectively), MSE (0.580 mg/day and 0.599 mg/day in internal and external validation, respectively), and RMSE (0.761 mg/day and 0.774 mg/day in internal and external validation, respectively) were achieved. Ideal predicted percentages were high in both internal (63.0%) and external validation (59.7%), respectively. Compared with the MLR model, the BPNN model showed a higher ideal prediction percentage in the external validation group (59.7% vs. 56.6%), and showed the best prediction accuracy in the intermediate-dose subgroup (internal validation group: 85.2%; external validation group: 84.7%) and a high predicted percentage in the high-dose subgroup (internal validation group: 36.2%; external validation group: 39.8%), but poor performance in the low-dose subgroup (internal validation group: 0%; external validation group: 0.3%). Meanwhile, the BPNN model showed better ideal prediction percentage in the high-dose group than the MLR model (internal validation: 36.2% vs. 31.6%; external validation: 42.8% vs. 37.8%). The BPNN model shows promise for predicting the warfarin maintenance dose after heart valve replacement.

The Prediction Model of Warfarin Individual Maintenance Dose for Patients Undergoing Heart Valve Replacement, Based on the Back Propagation Neural Network

Background: Warfarin acts in heart valve replacement patients to minimize thromboembolic complications. We investigated whether patients can be distinguished based on their genotypes to eff...

Genetic influence on bleeding and over-anticoagulation risk in patients undergoing warfarin treatment after heart valve replacements.

Thus far, the target value for international normalized ratio (INR) has remained to be determined. The current study aimed to further explore the INR value of the anti-coagulation drug warfarin after cardiac valve replacement. The clinical data of 213 patients who underwent cardiac valve replacement at Linyi Central Hospital (Linyi, China) between January 2010 and May 2013 were retrospectively analyzed. The warfarin dosage, prothrombin time (PT) and INR were compared among patients with hemorrhage or embolism, and those with no complications. A total of 31 cases (14.6%) developed adverse reactions and complications during the medication period, including 21 cases with hemorrhage (9.9%, hemorrhage group) and 10 cases with embolism (4.7%, embolism group), while 182 patients did not (85.4%, normal group). The average dosage of warfarin was 2.0±0.6, 3.1±0.7 and 1.7±0.6 mg/day in the normal, hemorrhage and embolism groups, respectively. The dosage of warfarin, the PT and the INR in the hemorrhage group were all significantly greater than those in the normal group and the embolism group (all P<0.05). INR monitoring is recommended to ensure the safety of the anti-coagulant drug warfarin, but further study is still required to determine a reasonable target INR value.

/pdf/international-normalized-ratio-for-the-guidance-of-warfarin-33qso7ww6b.pdf

International normalized ratio for the guidance of warfarin treatment in elderly patients after cardiac valve replacement

Background: Over the years, the internet has enabled considerable progress in the management of chronic diseases, especially hypertension and diabetes. It also provides novel opportunities in online anticoagulation management. Nevertheless, there is insufficient evidence regarding the effectiveness of online anticoagulation management.
Objective: This study explored the effectiveness and safety of warfarin management via the Alfalfa app, so as to provide evidence in support of anticoagulant management through online services.
Methods: In this retrospective, observational cohort study, 824 patients were included. In the offline group, patients went to the hospital clinic for warfarin management. In the Alfalfa app group, patients reported the dose of warfarin, current international normalized ratio (INR) value, and other related information through the Alfalfa app. Physicians or pharmacists used the app to adjust the dose of warfarin and determined the time for the next blood INR testing. Patients completed INR testing by point-of-care at home or hospital. The primary outcome of the study was the percentage of time in therapeutic range (TTR). Secondary outcomes included minor and major bleeding events, thrombotic events, warfarin-related emergency department visits, hospital admissions, and high INR values.
Results: The TTR and percentage of INR values in the range were significantly higher in the Alfalfa app group than in the offline group (79.35% vs 52.38%, P<.001; 3314/4282, 77.39% vs 2005/4202, 47.72%, P<.001, respectively). Patients managed via the Alfalfa app had lower rates of subtherapeutic (172/4282, 4.02% vs 388/4202, 9.23%; P<.001), supratherapeutic (487/4282, 11.37% vs 882/4202, 20.99%; P<.001), and extreme subtherapeutic INR values (290/4282, 6.77% vs 910/4202, 21.66%; P<.001). Additionally, the Alfalfa app group had lower incidences of major bleeding (2/425, 0.5% vs 12/399, 3.0%; P=.005), warfarin-related emergency department visits (13/425, 3.1% vs 37/399, 9.3%; P<.001), and hospital admissions (1/425, 0.2% vs 12/399, 3.0%; P=.001) compared with the offline group. However, the Alfalfa app group had a higher incidence of minor bleeding than the offline group (45/425, 10.6% vs 20/399, 5.0%; P=.003). There were similar incidences in extreme supratherapeutic INR values (19/4282, 0.44% vs 17/4202, 0.40%; P=.78) and thromboembolic events (1/425, 0.2% vs 1/399, 0.3%; P=.53) between the two groups.
Conclusions: Warfarin management is superior via the Alfalfa app than via offline services in terms of major bleeding events, warfarin-related emergency department visits, and hospital admissions.

/pdf/effectiveness-of-the-alfalfa-app-in-warfarin-therapy-3j67vdefze.pdf

Effectiveness of the Alfalfa App in Warfarin Therapy Management for Patients Undergoing Venous Thrombosis Prevention and Treatment: Cohort Study

Patients requiring low-dose warfarin are more likely to suffer bleeding due to overdose. The goal of this work is to improve the feedforward neural network model's precision in predicting the low maintenance dose for Chinese in the aspect of training data construction. We built the model from a resampled dataset created by equal stratified sampling (maintaining the same sample number in three dose-groups with a total of 3639) and performed internal and external validations. Comparing to the model trained from the raw dataset of 19,060 eligible cases, we improved the low-dose group's ideal prediction percentage from 0.7 to 9.6% and maintained the overall performance (76.4% vs. 75.6%) in external validation. We further built neural network models on single-dose subsets to invest whether the subsets samples were sufficient and whether the selected factors were appropriate. The training set sizes were 1340 and 1478 for the low and high dose subsets; the corresponding ideal prediction percentages were 70.2% and 75.1%. The training set size for the intermediate dose varied and was 1553, 6214, and 12,429; the corresponding ideal prediction percentages were 95.6, 95.1%, and 95.3%. Our conclusion is that equal stratified sampling can be a considerable alternative approach in training data construction to build drug dosing models in the clinic.

/pdf/warfarin-maintenance-dose-prediction-for-chinese-after-heart-qj1q98kuw1.pdf

Warfarin maintenance dose prediction for Chinese after heart valve replacement by a feedforward neural network with equal stratified sampling.

Objective
To investigate the optimal anticoagulation methods and monitoring strategy for Chinese patients undergoing heart valve replacement, which is potentially quite different from western populations.


Methods
In this multicenter prospective cohort study, the anticoagulation and monitoring strategy data was acquired from 25 773 in-hospital patients in 35 medical centers and 20 519 patients in outpatient clinic in 11 medical centers from January 1st, 2011 to December 31th, 2015.


Results
As for in-hospital patients, mean age of study population was (48.6±11.2) years old; main etiology of valve pathology was rheumatic (87.5%) origin among study cohort; 94.8% of study population received mechanical valve implantation; international normalized ratio (INR) monitoring (in all the study centers) and low-intensity anticoagulation strategy (31 hospitals chose target INR range of 1.5-2.5, and actual values of INR among 89.2% of 100 069 in-hospital monitoring samples were 1.5-2.5), with mean actual INR values of 1.84±0.53, and warfarin dosage of (2.82±0.93) mg/d were widely adopted among the study centers; strategies of in-hospital warfarin administration were similar in all the study centers; complication rates of low-intensity anticoagulation strategy were low in severe hemorrhage (0.02%), thrombosis (0.05%), and thromboembolism (0.05%) events, without anticoagulation-related death.As for 18 974 outpatient clinic patients, the follow-up rate was 92.47%, with a total of 30 012 patient-years (Pty). Anticoagulation-related morbidity and mortality rates were 0.67% and 0.15% Pty; major hemorrhage morbidity and mortality rates were 0.25% and 0.13% Pty; thromboembolism morbidity and mortality rates were 0.45% and 0.03% Pty.The mean dosage of warfarin daily dosage was (2.85±1.23) mg/d and INR value was 1.82±0.57.No significant regional difference in the intensity of anticoagulation therapy was noted during the study.


Conclusions
INR can be used as a normalized indicator for intensity of anticoagulation therapy in China.The optimal anticoagulation intensity with INR range from 1.5 to 2.5 is safe and effective for Chinese patients with heart valve replacement, and there is no significant regional difference in the intensity of anticoagulation therapy.


Key words: 
Heart valve prosthesis implantation; Low-intensity anticoagulation therapy; Multicenter study; Follow-up studies

Zheng Xm

Papers

[The multicenter study on the registration and follow-up of low anticoagulation therapy for the heart valve operation in China].