International Warfarin Pharmacogenetics Consortium Research Articles

Meta-analysis of genome-wide association studies of stable warfarin dose in patients of African ancestry

AbstractWarfarin dose requirements are highly variable because of clinical and genetic factors. Although genetic variants influencing warfarin dose have been identified in European and East Asian populations, more work is needed to identify African-specific genetic variants to help optimize warfarin dosing. We performed genome-wide association studies (GWASs) in 4 African cohorts from Uganda, South Africa, and Zimbabwe, totaling 989 warfarin-treated participants who reached stable dose and had international normalized ratios within therapeutic ranges. We also included 2 African American cohorts recruited by the International Warfarin Pharmacogenetics Consortium (n = 316) and the University of Alabama at Birmingham (n = 199). After the GWAS, we performed standard error-weighted meta-analyses and then conducted stepwise conditional analyses to account for known loci in chromosomes 10 and 16. The genome-wide significance threshold was set at P < 5 × 10−8. The meta-analysis, comprising 1504 participants, identified 242 significant SNPs across 3 genomic loci, with 99.6% of these located within known loci on chromosomes 10 (top SNP: rs58800757, P = 4.27 × 10−13) and 16 (top SNP: rs9925964, P = 9.97 × 10−16). Adjustment for the VKORC1 SNP -1639G>A revealed an additional locus on chromosome 2 (top SNPs rs116057875/rs115254730/rs115240773, P = 3.64 × 10−8), implicating the MALL gene, that could indirectly influence warfarin response through interactions with caveolin-1. In conclusion, we reaffirmed the importance of CYP2C9 and VKORC1 in influencing warfarin dose requirements, and identified a new locus (MALL), that still requires direct evidence of biological plausibility.

Validating two international warfarin pharmacogenetic dosing algorithms for estimating the maintenance dose for patients in Singapore

Predicting optimal warfarin dosing is difficult due to complex pharmacodynamics and pharmacokinetics, narrow therapeutic index and susceptibility to many factors.1 Genetic variations of the CYP2C9 and VKORC1 enzymes, occurring in different frequencies in different populations, play a significant role in determining warfarin dosing.1-4 Using pharmacogenetic dosing algorithms to predict warfarin doses may shorten the time to achieve target International Normalised Ratio (INR) and stable dose.2,5 The Clinical Pharmacogenetics Implementation Consortium Guidelines 2017 Update4 recommends the Gage (WarfarinDosing.org7) and International Warfarin Pharmacogenetics Consortium (IWPC)8 pharmacogenetic algorithms.

Genomewide Association Study Identifies Copy Number Variants Associated With Warfarin Dose Response and Risk of Venous Thromboembolism in African Americans.

The anticoagulant warfarin is commonly used to control and prevent thrombotic disorders, such as venous thromboembolism (VTE), which disproportionately afflicts African Americans. Despite the importance of copy number variants (CNVs), few studies have focused on characterizing and understanding their role in drug response and disease risk among African Americans. In this study, we conduct the first genome-wide analysis of CNVs to more comprehensively account for the contribution of genetic variation in warfarin dose requirement and VTE risk among African Americans. We used hidden Markov models to detect CNVs from high-throughput single-nucleotide polymorphism arrays for 340 African American participants in the International Warfarin Pharmacogenetics Consortium. We identified 11,570 CNVs resulting in 2,038 copy number variable regions (CNVRs) and found 3 CNVRs associated with warfarin dose requirement and 3 CNVRs associated with VTE risk in African Americans. CNVRs 1q31.2del and 6q14.1del were associated with increased warfarin dose requirement (β=11.18 and 4.94, respectively; Pemp =< 0.002); CNVR 19p13.31del was associated with decreased warfarin dose requirement (β=-1.41, Pemp =0.0004); CNVRs (2p22.1del and 5q35.1-q35.2del) were found to be associated with increased risk of VTE (odds ratios (ORs)=1.88 and 14.9, respectively; Pemp ≤0.02); and CNVR 10q26.12del was associated with a decreased risk of VTE (OR=0.6; Pemp =0.05). Modeling of the 10q26.12del in HepG2 cells revealed that this deletion results in decreased fibrinogen gene expression, decreased fibrinogen and WDR11 protein levels, and decreased secretion of fibrinogen into the extracellular matrix. We found robust evidence that CNVRs could contribute to warfarin dose requirement and risk of VTE in African Americans and for 10q26.3del describe a possible pathogenic mechanism.

Comparison of Maintenance Dose Predictions by Warfarin Dosing Algorithms Based on Chinese and Western Patients.

Warfarin has a long record of safe and effective clinical use, and it remains one of the most commonly prescribed drugs for the prevention and treatment of thromboembolic conditions even in the era of direct oral anticoagulants. To address its large interindividual variability and narrow therapeutic window, the Clinical Pharmacogenetics Implementation Consortium has recommended using pharmacogenetic dosing algorithms, such as the ones developed by the International Warfarin Pharmacogenetics Consortium (IWPC) and by Gage et al, to dose warfarin when genotype information is available. In China, dosing algorithms based on local patient populations have been developed and evaluated for predictive accuracy of warfarin maintenance doses. In this study, percentage deviations of doses predicted by 15 Chinese dosing algorithms from that by IWPC and Gage algorithms were systematically evaluated to understand the differences between Chinese and Western algorithms. In general, dose predictions by Chinese dosing algorithms tended to be lower than those predicted by IWPC or Gage algorithms for the most prevalent VKORC1 and CYP2C9 genotypes in the Chinese population. The extent of negative prediction deviation appeared to be largest in the younger age group with smaller body weight. Our findings are consistent with previous reports that Asians have a higher sensitivity to warfarin and require lower doses than Western populations.

Warfarin hypersensitivity in the internist's practice: a case report

Anticoagulants are widely used in clinical practice to reduce the risk of cardiovascular events. However, there are associated clinical conditions that require coagulation system monitoring due to an increased risk of bleeding. In clinical practice, cases of warfarin hypersensitivity due to gene polymorphisms are known. Taking warfarin in such a situation is often manifested by massive bleeding that threatens the patient's life. Sixty seven-years-old female patient was admitted to the internal medicine department of Rostov State Medical University clinic in March 2022. The day before, she noted severe abdominal pain, loose bloody stool. An outpatient examination revealed an increase in international normalized ratio to 9,42, thrombin time &gt;30 sec. There were no signs of primary gastrointestinal pathology. According to anamnesis, on January 21, 2022, the patient underwent mechanical aortic valve replacement, in connection with which warfarin was prescribed at a dose of 2,5 mg/day. After establishing hypocoagulation, warfarin was discontinued. A pharmacogenetic analysis was performed, which established the carriage of polymorphisms of cytochrome P450 system genes, the homozygous mutation 1075A&gt;C (CYP2C9 (*3/*3)), and the heterozygous mutation of the vitamin K reductase gene VKORC1 G(-1639)A (VKORC1 GA). The individual dosage of warfarin was calculated according to International Warfarin Pharmacogenetics Consortium guidelines, which was 9 mg/week. After adjusting the dose of warfarin, the level of international normalized ratio decreased to 3,61, thrombin time to 13 sec. The patient was discharged with recommendations to follow an individual warfarin regimen and monitor coagulation parameters. The presence of hypersensitivity to warfarin is not a reason for its complete withdrawal in situations requiring longterm anticoagulation. In this regard, it is necessary for the doctor to be vigilant when prescribing warfarin to a patient, understanding the causes and methods for diagnosing hypersensitivity to warfarin, and timely correction of its dosage.

Abstract 10926: Warfarin Dosing Algorithms With the Information on Genotype and Amiodarone for Patients Without Extracorporeal Circulation Devices Could Be Beneficial for Those With Left Ventricular Assist Devices

Introduction: Many prediction algorithms for warfarin maintenance dose (WD) were constructed based on the clinical and genetic information of the patients without extracorporeal circulation devices (ECDs). However, it is unclear whether might be useful for the patients with ECD, such as left ventricular assist devices (LVAD), who still require warfarin. Hypothesis: The WD algorithms constructed using non-ECD patients could be useful for LVAD patients. Methods: An observational study, the Human Genome Research Ethics Committee of OU approved (#756), was conducted at the Osaka University (OU) Hospital on 108 LVAD patients receiving warfarin therapy. Genetic polymorphisms of CYP2C9*1, *3, and VKORC1-1639G&gt;A were tested using TaqMan genotyping assay kits. WD was defined as the administered dose during the periods when the PT-INR value was within its target range. We investigated the difference between the actual warfarin dose (AWD) and the calculated warfarin dose (CWD) using an algorithm proposed by the International Warfarin Pharmacogenetics Consortium (IWPC), which we verified with 125 Japanese non-ECD patients previously (Eur J Clin Pharmacol, 75:901). Results: The percentage of patients whose difference between CWD and AWD was within 20% of AWD showed no significant difference between the LVAD and the non-ECD (%; I, LVAD, 41.7, non-ECD, 49.6, p=0.23). The root mean squared percentage error (RMSPE) to the AWD was similar between the groups (%; LVAD, 42, non-ECD, 37). As the algorithm requires if amiodarone is concomitant with warfarin, we probed the effect of the amiodarone dose on the CWD. Interestingly, RMSPE of the LVAD with ≥ 200 mg/day of amiodarone (≥ 200A) was higher compared to those with &lt; 200 mg/day (&lt;200A) and without amiodarone (%; 85.5, 25.9, 36.8). The difference between AWD and CWD of ≥ 200A was significantly higher than &lt;200A (mg/day; 0.87, -0.07; p=0.012). We verified similar tendencies using other algorithms published before. Conclusions: The WD algorithm with the information on genotype and concomitant use of amiodarone for the non-ECD could be beneficial for the LVAD. Besides them, extra careful attention should be required for patients with ≥ 200 mg/day of amiodarone not to have warfarin overdosed.

Validation and comparison between two warfarin dosing clinical algorithms and warfarin fixed dosing in specialized heart center: cross-sectional study.

Warfarin is well known as a narrow therapeutic index that has prodigious variability in response which challenges dosing adjustment for the maintenance of therapeutic international normalized ratio. However, an appreciated population not on new oral anticoagulants may still need to be stabilized with warfarin dosing. The current study's main objective was to validate and compare two models of warfarin clinical algorithm models namely the Gage and the International Warfarin Pharmacogenetics Consortium (IWPC) with warfarin 5 mg fixed standard dosing strategy in a sample of Sudanese subjects. We have conducted a cross-sectional study recruited from the out-patient clinic at a tertiary specialized heart center. We included subjects with unchanged warfarin dose (stabilized), and with therapeutic international normalized ratio. The predicted doses of warfarin in the two models were calculated by three different methods (accuracy, clinical practicality, and the clinical safety of the clinical algorithms). The primary outcomes were the measurements of the clinical (accuracy, practicality, and safety) in each of the two clinical algorithms models compared to warfarin 5 mg fixed standard dose strategy. We have enrolled 71 Sudanese subjects with mean age (51.7 ± 14 years), of which (49, 69.0%) were females. There was no significant difference between the warfarin 5 mg fixed standard dose strategy and the predicted doses of the two clinical algorithm models (MAE 1.44, 1.45, and 1.49 mg/day [P =0.4]) respectively. In the clinical practicality, all of the three models had a high percent of subjects (95.0%, 51.9%, and 66.7%) in the ideal dose range in middle dose group (3-7 mg/ day) for warfarin 5 mg fixed standard dosing strategy, Gage, and IWPC clinical algorithm models respectively. However, a small percent of subjects was exhibited in the warfarin low dose group ≤ 3 mg/day (0.0%, 15.0%, and 10.0%) and warfarin high dose group ≥ 7 mg/day (0.0%, 33.3%, and 33.3%) for warfarin 5 mg fixed standard dosing strategy, Gage, and IWPC clinical algorithms respectively. In terms of clinical safety, the percent of subjects with severely over-prediction were 28.2%, 22.5%, and 22.5% for warfarin 5 mg fixed standard dosing, Gage, and IWPC, respectively. While the percent of severely under-prediction was 12.7%, 7.0%, and 5.6% for the warfarin 5 mg fixed standard dosing, Gage, and IWPC, respectively. The Gage and IWPC clinical algorithm models were accurate, more clinically practical, and clinically safe than warfarin 5 mg standard dosing in the study population. The cardiologist can use either models (Gage and IWPC) to stratify subjects for accurate, practical, and clinically safe warfarin dosing..

Warfarin sensitivity is associated with increased hospital mortality in critically Ill patients.

BackgroundWarfarin is a widely used anticoagulant with a narrow therapeutic index and large interpatient variability in the therapeutic dose. Warfarin sensitivity has been reported to be associated with increased incidence of international normalized ratio (INR) > 5. However, whether warfarin sensitivity is a risk factor for adverse outcomes in critically ill patients remains unknown. In the present study, we aimed to evaluate the utility of different machine learning algorithms for the prediction of warfarin sensitivity and to determine the impact of warfarin sensitivity on outcomes in critically ill patients.MethodsNine different machine learning algorithms for the prediction of warfarin sensitivity were tested in the International Warfarin Pharmacogenetic Consortium cohort and Easton cohort. Furthermore, a total of 7,647 critically ill patients was analyzed for warfarin sensitivity on in-hospital mortality by multivariable regression. Covariates that potentially confound the association were further adjusted using propensity score matching or inverse probability of treatment weighting.ResultsWe found that logistic regression (AUC = 0.879, 95% CI: 0.834–0.924) was indistinguishable from support vector machine with a linear kernel, neural network, AdaBoost and light gradient boosting trees, and significantly outperformed all the other machine learning algorithms. Furthermore, we found that warfarin sensitivity predicted by the logistic regression model was significantly associated with worse in-hospital mortality in critically ill patients with an odds ratio (OR) of 1.33 (95% CI, 1.01–1.77).ConclusionsOur data suggest that the logistic regression model is the best model for the prediction of warfarin sensitivity clinically and that warfarin sensitivity is likely to be a risk factor for adverse outcomes in critically ill patients.

Building and analyzing machine learning-based warfarin dose prediction models using scikit-learn.

For personalized drug dosing, prediction models may be utilized to overcome the inter-individual variability. Multiple linear regression has been used as a conventional method to model the relationship between patient features and optimal drug dose. However, linear regression cannot capture non-linear relationships and may be adversely affected by non-normal distribution and collinearity of data. To overcome this hurdle, machine learning models have been extensively adapted in drug dose prediction. In this tutorial, random forest and neural network models will be trained in tandem with a multiple linear regression model on the International Warfarin Pharmacogenetics Consortium dataset using the scikit-learn python library. Subsequent model analyses including performance comparison, permutation feature importance computation and partial dependence plotting will be demonstrated. The basic methods of model training and analysis discussed in this article may be implemented in drug dose-related studies.

Machine Learning for Prediction of Stable Warfarin Dose in US Latinos and Latin Americans.

Populations used to create warfarin dose prediction algorithms largely lacked participants reporting Hispanic or Latino ethnicity. While previous research suggests nonlinear modeling improves warfarin dose prediction, this research has mainly focused on populations with primarily European ancestry. We compare the accuracy of stable warfarin dose prediction using linear and nonlinear machine learning models in a large cohort enriched for US Latinos and Latin Americans (ULLA). Each model was tested using the same variables as published by the International Warfarin Pharmacogenetics Consortium (IWPC) and using an expanded set of variables including ethnicity and warfarin indication. We utilized a multiple linear regression model and three nonlinear regression models: Bayesian Additive Regression Trees, Multivariate Adaptive Regression Splines, and Support Vector Regression. We compared each model’s ability to predict stable warfarin dose within 20% of actual stable dose, confirming trained models in a 30% testing dataset with 100 rounds of resampling. In all patients (n = 7,030), inclusion of additional predictor variables led to a small but significant improvement in prediction of dose relative to the IWPC algorithm (47.8 versus 46.7% in IWPC, p = 1.43 × 10−15). Nonlinear models using IWPC variables did not significantly improve prediction of dose over the linear IWPC algorithm. In ULLA patients alone (n = 1,734), IWPC performed similarly to all other linear and nonlinear pharmacogenetic algorithms. Our results reinforce the validity of IWPC in a large, ethnically diverse population and suggest that additional variables that capture warfarin dose variability may improve warfarin dose prediction algorithms.

Performance Evaluation of Warfarin Dose Prediction Algorithms and Effects of Clinical Factors on Warfarin Dose in Chinese Patients.

The clinical utility of warfarin dose prediction algorithms remains controversial, our purpose is to evaluate the performance of warfarin dose prediction algorithms and the effects of clinical factors on warfarin dose in Chinese patients. Clinical data of 217 patients who received warfarin treatment were used to assess 6 warfarin dose prediction algorithms (OHNO, IWPC [International Warfarin Pharmacogenetics Consortium], HUANG, KIM, BRESS, and MIAO). The predicted dose (PD) was compared with the warfarin optimal dose (WOD, defined as the dose that maintains the international normalized ratio within the target range of 2.0-3.0). A multiple regression analysis with WOD as the dependent variable was performed to evaluate the effects of clinical factors on warfarin dose. The mean absolute error analysis ranked the predictive accuracies of the algorithms as OHNO > IWPC > HUANG > KIM > BRESS > MIAO. Stratified analysis indicated that HUANG most accurately predicted that patients required lower WODs (≤3 mg/d), whereas OHNO was the most effective in predicting medium WODs (3-5 mg/d). KIM was effective in predicting high WODs (>5 mg/d). Multiple linear regression analysis showed that VKORC1 (rs9923231) and body mass index were significantly positively correlated with WOD, whereas concurrent atrial fibrillation status, CYP2C9*3 (rs1057910), and sex were significantly negatively correlated with WOD. In Chinese patients, OHNO should be given priority during the prediction and selection of warfarin dose. When using OHNO to predict warfarin dose (≤3 mg/d or >5 mg/d), HUANG or KIM algorithms can provide precise predictions. At the same time, physicians should pay close attention to clinical factors, such as VKORC1 (rs9923231), concurrent atrial fibrillation status, CYP2C9*3 (rs1057910), body mass index, and sex, to improve warfarin dose adjustment strategies in Chinese patients.

Evolutionary synthetic oversampling technique and cocktail ensemble model for warfarin dose prediction with imbalanced data

To improve the accuracy of warfarin daily dose prediction, we develop an evolutionary synthetic oversampling technique (ESMOTE) with a cocktail ensemble model (CEM) called ESMOTE-CEM. Different from conventional oversampling methods, ESMOTE finds the near-optimal oversampling parameters by evolving the parameter representation based on the pre-predicted warfarin dose and then synthesizes new samples to balance the data. The CEM, which improves the performance of random forest (RF) and boosted regression tree (BRT) models using a hybrid mechanism in the regression calculation, estimates the daily dose of warfarin. We test the ESMOTE-CEM on a dataset of 733 samples collected from the First Affiliated Hospital of Soochow University and the International Warfarin Pharmacogenetics Consortium (IWPC). The results show that ESMOTE outperformed the other oversampling methods by at least 6.98% for R2 and 5.03% for the mean squared error (MSE). In terms of the percentage of patients whose predicted warfarin dose is within 20% of the actual stable therapeutic dose (20%-p value), the ESMOTE-CEM achieves a 20%-p value of 50%. Moreover, compared to RF, BRT and AdaBoost models, the CEM is the most suitable base predictive model for ESMOTE.

Developing and Validating a Clinical Warfarin Dose-Initiation Model for Black-African Patients in South Africa and Uganda.

Warfarin remains the oral anticoagulant of choice in sub-Saharan Africa. However, dosing is challenging due to a highly variable clinical response for a given dose. This study aimed to develop and validate a clinical warfarin dose-initiation model in sub-Saharan Black-African patients. For the development cohort, we used data from 364 patients who were recruited from 8 outpatient clinics and hospital departments in Uganda and South Africa (June 2018-July 2019). Validation was undertaken using the International Warfarin Pharmacogenetics Consortium (IWPC) dataset (690 black patients). Four predictors (age, weight, target International Normalized Ratio range, and HIV status) were included in the final model, which achieved mean absolute errors (MAEs; mean of absolute differences between true dose and dose predicted by the model) of 11.6 (95% confidence interval (CI) 10.4-12.8) and 12.5 (95% CI 11.6-13.4) mg/week in the development and validation cohorts, respectively. Two other clinical models, IWPC and Gage, respectively, obtained MAEs of 12.5 (95% CI 11.3-13.7) and 12.7 (95% CI 11.5-13.8) mg/week in the development cohort, and 12.1 (95% CI 11.2-13.0) and 12.2 (95% CI 11.4-13.1) mg/week in the validation cohort. Compared with fixed dose-initiation, our model decreased the percentage of patients at high risk of suboptimal anticoagulation by 7.5% (1.5-13.7%) and 11.9% (7.1-16.8%) in the development and validation cohorts, respectively. The clinical utility of this model will be tested in a prospective study. Study Highlights WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC? ☑ Warfarin dosing remains challenging due to a highly variable clinical response for a given dose. WHAT QUESTION DID THIS STUDY ADDRESS? ☑ Can a clinical dose-initiation model be developed and validated for sub-Saharan Black-African patients? WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE? ☑ We have developed the first warfarin dose-initiation clinical model for Black-African patients in Uganda and South Africa. HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE? ☑ We will be implementing and validating this model in a prospective cohort to inform future large-scale implementation. More optimized dosing should improve the quality of warfarin anticoagulation in these two developing countries.

Differences in Predicted Warfarin Dosing Requirements Between Hmong and East Asians Using Genotype-Based Dosing Algorithms.

Warfarin's narrow therapeutic index and high variability in dosage requirements make dosage selection critical. Genetic factors are known to impact warfarin dosage selection. The Hmong are a unique Asian subpopulation numbering over 278,000 in the United States whose participation in genetics-based research is virtually nonexistent. The translational significance of early reports of warfarin pharmacogene differences in Hmong has not been evaluated. (i) To validate previously identified allele frequency differences relevant to warfarin dosing in Hmong versus East Asians and (ii) to compare predicted warfarin sensitivity and maintenance doses between a Hmong population and an East Asian cohort. DNA collected from two independent cohorts (n=236 and n=198) of Hmong adults were genotyped for CYP2C9 (*2, *3), VKORC1 (G-1639A), and CYP4F2 (*3). Allele frequencies between the combined Hmong cohort (n=433) and East Asians (n=1165) from the 2009 International Warfarin Pharmacogenetics Consortium (IWPC) study were compared using a χ2 test. Percentages of Hmong and East Asian participants predicted to be very sensitive to warfarin were compared using a χ2 test, and the predicted mean warfarin maintenance dose was compared with a t test. The allele frequencies of CYP2C9*3 in the combined Hmong cohort and CYP4F2*3 in the VIP-Hmong cohort are significantly different from those in East Asians (18.9% vs 3.0%, p<0.001 and 9.8% vs 22.1%, p<0.001, respectively). Comparing the combined Hmong cohort to the East Asian cohort, the percentage of participants predicted to be very sensitive to warfarin was significantly higher (28% vs 5%, p<0.01) and the mean predicted warfarin maintenance dose was significantly lower (19.8 vs 21.3mg/week, p<0.001), respectively. The unique allele frequencies related to warfarin when combined with nongenetic factors observed in the Hmong translate into clinically relevant differences in predicted maintenance dose requirements for Hmong versus East Asians.

Evaluating warfarin dosing models on multiple datasets with a novel software framework and evolutionary optimisation

Warfarin is an effective preventative treatment for arterial and venous thromboembolism, but requires individualised dosing due to its narrow therapeutic range and high individual variation. Many machine learning techniques have been demonstrated in this domain. This study evaluated the accuracy of the most promising algorithms on the International Warfarin Pharmacogenetics Consortium dataset and a novel clinical dataset of South African patients. Support vectors and linear regression were amongst the top performers in both datasets and performed comparably to recent stacked ensemble approaches, whilst neural networks were one of the worst performers in both datasets. We also introduced genetic programming to automatically optimise model architectures and hyperparameters without human guidance. Remarkably, the generated models were found to match the performance of the best models hand-crafted by human experts. Finally, we present a novel software framework (Warfit-learn) for warfarin dosing research. It leverages the most successful techniques in preprocessing, imputation, and parallel evaluation, with the goal of accelerating research and making results in this domain more reproducible.

DBCSMOTE: a clustering-based oversampling technique for data-imbalanced warfarin dose prediction

BackgroundVitamin K antagonist (warfarin) is the most classical and widely used oral anticoagulant with assuring anticoagulant effect, wide clinical indications and low price. Warfarin dosage requirements of different patients vary largely. For warfarin daily dosage prediction, the data imbalance in dataset leads to inaccurate prediction on the patients of rare genotype, who usually have large stable dosage requirement. To balance the dataset of patients treated with warfarin and improve the predictive accuracy, an appropriate partition of majority and minority groups, together with an oversampling method, is required.MethodTo solve the data-imbalance problem mentioned above, we developed a clustering-based oversampling technique denoted as DBCSMOTE, which combines density-based spatial clustering of application with noise (DBCSCAN) and synthetic minority oversampling technique (SMOTE). DBCSMOTE automatically finds the minority groups by acquiring the association between samples in terms of the clinical features/genotypes and the warfarin dosage, and creates an extended dataset by adding the new synthetic samples of majority and minority groups. Meanwhile, two ensemble models, boosted regression tree (BRT) and random forest (RF), which are built on the extended dataset generateed by DBCSMOTE, accomplish the task of warfarin daily dosage prediction.ResultsDBCSMOTE and the comparison methods were tested on the datasets derived from our Hospital and International Warfarin Pharmacogenetics Consortium (IWPC). As the results, DBCSMOTE-BRT obtained the highest R-squared (R2) of 0.424 and the smallest mean squared error (mse) of 1.08. In terms of the percentage of patients whose predicted dose of warfarin is within 20% of the actual stable therapeutic dose (20%-p), DBCSMOTE-BRT can achieve the largest value of 47.8% among predictive models. The more important thing is that DBCSMOTE saved about 68% computational time to achieve the same or better performance than the Evolutionary SMOTE, which was the best oversampling method in warfarin dose prediction by far. Meanwhile, in warfarin dose prediction, it is discovered that DBCSMOTE is more effective in integrating BRT than RF for warfarin dose prediction.ConclusionOur finding is that the genotypes, CYP2C9 and VKORC1, no doubt contribute to the predictive accuracy. It was also discovered left atrium diameter, glutamic pyruvic transaminase and serum creatinine included in the model actually improved the predictive accuracy; When congestive heart failure, diabetes mellitus and valve replacement were absent in DBCSMOTE-BRT/RF, the predictive accuracy of DBCSMOTE-BRT/RF decreased. The oversampling ratio and number of minority clusters have a large impact on the effect of oversampling. According to our test, the predictive accuracy was high when the number of minority clusters was 6 ~ 8. The oversampling ratio for small minority clusters should be large (> 1.2) and for large minority clusters should be small (< 0.2). If the dataset becomes larger, the DBCSMOTE would be re-optimized and its BRT/RF model should be re-trained. DBCSMOTE-BRT/RF outperformed the current commonly-used tool called Warfarindosing. As compared to Evolutionary SMOTE-BRT and RF models, DBCSMOTE-BRT and RF models take only a small computational time to achieve the same or higher performance in many cases. In terms of predictive accuracy, RF is not as good as BRT. However, RF still has a powerful ability in generating a highly accurate model as the dataset increases; the software “WarfarinSeer v2.0” is a test version, which packed DBCSMOTE-BRT/RF. It could be a convenient tool for clinical application in warfarin treatment.

Evaluation of Stable Doses of Warfarin in a Patient Cohort.

Definitions for stable dose of warfarin varies in the reported studies. International warfarin pharmacogenetic consortium (IWPC) algorithm was generated from the data based on these definitions. In the present study, we primarily evaluated whether any significant differences exist between the definitions for stable warfarin dose. A prospective cross-sectional study in adults receiving warfarin for at least 3 months was carried out. Stable doses of warfarin as defined in previous studies were compared with the standard definition. Bland-Altman plots, Pearson's correlation and intra-class coefficients (ICC) were used to assess the correlation, reliability and agreements between the doses. Sixty-four patients were recruited. Twenty definitions were obtained from the previous studies. We observed that all but one showed very high or high positive correlations; and either excellent or good ICC. No significant differences between the doses initiated and predicted by IWPC algorithm. We observed similar stable doses between the definitions except for one. Hence, IWPC algorithm may not have any bias associated with inclusion of any studies with variable definitions for stable warfarin dose.

Pharmacogenetic-Guided Algorithm to Improve Daily Dose of Warfarin in Elder Han-Chinese Population.

ObjectivesTo verify the accuracy of the International Warfarin Pharmacogenetics Consortium (IWPC) algorithm, identify the effects of genetic and clinical factors on warfarin stable dose, and to establish a new warfarin stable dose prediction algorithm for the elderly Han-Chinese population under the guidance of pharmacogenetics.MethodsAccording to the inclusion criteria, 544 non-valvular atrial fibrillation patients taking warfarin for anticoagulation treatment were enrolled. Data information of three groups including the whole population, people under 65 years old and over 65 years old were substituted into the IWPC algorithm respectively to verify its accuracy. The basic data and clinical information of 360 elderly people were collected for statistical analysis and the genotypes of VKORC1-G1639A and CYP2C9 were detected by Sanger sequencing. The new algorithm of the elder pharmacogenetics warfarin dosing was obtained by stepwise multiple regression. The determination coefficient (R2), root mean squared error (RMSE), and the proportion of the predicted value within the true value range of ±20%(20%-p) were used to evaluate the accuracy of the IWPC algorithm and the new algorithm.ResultsAmong the three different age groups, the warfarin stable dose predictive accuracy of IWPC algorithm was the lowest in the elderly patients above 65-year-old. In this study, the important factors influencing the stable dose of warfarin in the elderly Han-Chinese were height, weight, body surface area, serum creatinine level, amiodarone usage, CYP2C9 (*1*2, *1*3), and VKORC1 (GG/GA) genotypes. By means of stepwise multiple regression analysis, we established a new elder warfarin dosing algorithm (R2=0.3714) containing height, creatinine, amiodarone usage, CYP2C9 (*1*2 or *1*3), and VKORC1 (GA or GG) genotypes. The prediction accuracy and clinical availability of the Elderly algorithm was significantly better than that of IWPC algorithm verified by RMSE, R2, and (20%-p) methods.ConclusionsThe IWPC model may not be suitable for the elder Han-Chinese population. Polymorphism of CYP2C9 and VKORC1 obviously affected warfarin stable dose of the elder Han-Chinese. Combination of genetic data with demographic and clinical factors could help to better improve warfarin doses in the elder Han-Chinese population.

Evaluation of the impact of body mass index on venous thromboembolism risk factors.

In this paper, we investigate the interaction impacts of body mass index (BMI) on the other important risk factors for venous thromboembolism (VTE), using deep venous thrombosis (DVT) patient data from the International Warfarin Pharmacogenetics Consortium (IWPC). We apply eight machine learning techniques, including naive Bayes classifier (NB), support vector machine (SVM), elastic net regression (ENET), logistic regression (LR), lasso regression (LAR), multivariate adaptive regression splines (MARS), boosted regression tree (BRT) and random forest model (RF). The RF method is selected as the best model for classification. Out of 33 features considered in this study, we identify 12 variables as relatively important risk factors for VTE. Finally, we examine the interaction impacts of BMI on these important VTE risk factors. We conclude that the impacts of risk factors on VTE incidence are varying across different BMI groups, and the variations are different for different risk factors. Therefore the interaction impacts of BMI on the other risk factors have to be taken into account in order to better understand the incidence of VTE.

Pharmacogenetic-guided and clinical warfarin dosing algorithm assessments with bleeding outcomes risk-stratified by genetic and covariate subgroups

BackgroundSafe administration of warfarin presents challenges due to a narrow therapeutic INR range and significant variability in inter-individual dose response. Bleeding secondary to warfarin use is a leading cause of hospitalization. MethodsFive warfarin dosing algorithms were assessed for accuracy of predicted compared to the INR target dose for patients with a HAS-BLED score ≥3 participating in the ENGAGE-AF TIMI 48 trial. Three warfarin metabolism subgroups (normal, sensitive, and highly sensitive responders) were established based on genotype. Mean differences between calculated and prescribed dose were determined for each algorithm across groups. ResultsA total of 7036 patients were analyzed and 1846 participants with HAS-BLED ≥3 were genotyped. The mean absolute error of predicted versus INR target dose for warfarin ranged from 8.1 mg/week on the pharmacogenetic-guided International Warfarin Pharmacogenetics Consortium (IWPC) and Gage algorithms to 11.3 mg/week on fixed dose. Overestimation of INR target dose occurred in 98% of highly sensitive responders by 21 mg/week for subjects on fixed dose. Pharmacogenetic-guided IWPC saw 89% overestimation with mean difference of 8.3 mg/week for highly sensitive responders. Major or clinically relevant non-major bleeding in the first 90 days of beginning warfarin was 3.27 times more likely for highly sensitive than normal responders. ConclusionsInitial doses were higher than INR target doses in high-risk bleeding subpopulations as defined by the modified HAS-BLED and genotype sensitivity analysis when compared to established algorithms. Clinical and pharmacogenetic-guided algorithms improved dosing in highly sensitive responders with HAS-BLED ≥3 compared to fixed dosing.