Estimation of the warfarin dose with clinical and pharmacogenetic data

Comparative performance of pharmacogenetics-based warfarin dosing algorithms in Chinese population: use of a pharmacokinetic/pharmacodynamic model to explore dosing regimen through clinical trial simulation

OBJECTIVE

Warfarin has a narrow therapeutic window and large variability in dosing that are affected by clinical and genetic factors. To help guide the dosing of warfarin, the Clinical Pharmacogenetics Implementation Consortium has recommended the use of pharmacogenetic algorithms, such as the ones developed by the International Warfarin Pharmacogenetics Consortium (IWPC) and by Gage et al. when genotype information is available.

METHODS

In this study, simulations were performed in Chinese cohorts to explore how dosing differences between Western (by IWPC and Gage et al.) and Chinese algorithms (by Miao et al.) would mean in terms of anticoagulation effect in clinical trials. We first tried to replicate a published clinical trial comparing genotype-guided dosing to routine clinical dosing in Chinese patients. We then made simulations where Chinese cohorts received daily doses recommended by Gage, IWPC, and Miao algorithms.

RESULTS

We found that in simulation conditions where dosing specifications were strictly followed, genotype-guided dosing by IWPC and Lenzini formulae was more likely to overshoot the upper limit of the therapeutic window by day 15, and thus may have a lower % time in therapeutic range (%TTR) than that of clinical dosing group. Also, in comparing Gage, IWPC, and Miao algorithms, we found that the Miao dosing cohort has the highest %TTR and the lowest risk of over-anticoagulation by day 28.

CONCLUSION

In summary, our results confirmed that algorithms developed based on data from local patients may be more suitable for achieving therapeutic international normalized ratio window in Chinese population.

Accuracy of an internationally validated genetic-guided warfarin dosing algorithm compared to a clinical algorithm in an Arab population

PURPOSE

To identify the impact of CYP2C9*2, *3, VKORC1-1639 G>A and CYP4F2*3 on warfarin dose in an Arab population. To compare the accuracy of a clinical warfarin dosing (CWD) versus genetic warfarin dosing algorithms (GWD) during warfarin initiation.

METHODS

A cohort of Arab patients newly starting on warfarin had their dose calculated using CWD published in www.warfarindosing.org and were followed for 1 month. Each patient provided a saliva sample. DNA was extracted, purified and genotyped for VKORC-1639 G>A, CYP2C9*2, CYP2C9*3 and CYP4F2*3. After reaching warfarin maintenance dose, the dose was recalculated using the GWD and median absolute error (MAE) and the percentage of warfarin doses within 20% of the actual dose were calculated and compared for the two algorithms.

RESULTS

The study enrolled 130 patients from 12 Arabian countries. Compared to those with wild type, carriers of reduced function alleles in CYP2C9 required significantly lower median (IQR) warfarin weekly dose [24.5 (15.3) vs. 35 (29.8) mg/week, p=0.006]. With regards to VKORC, patients with AA genotype had a significantly lower median (IQR) weekly warfarin dose compared to AG and GG [21(10.5) vs 29.4 (21), p<0.001 for AA vs AG, p<0.001 for AA vs GG]. The MAE (IQR) for the weekly dose of the GWD was significantly lower compared to CWD [8.1 (10.5) vs 12.4 (12.6) (p<0.001)].

CONCLUSION

CYP2C9 and VKORC1 variants are important determinants of warfarin dose in the Arab population. The use of the genetic and clinical factors led to better warfarin dose estimation when compared to clinical factors alone.

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

ABSTRACT

Warfarin 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.

Clinical decision making under uncertainty: a bootstrapped counterfactual inference approach

BACKGROUND

Learning policies for decision-making, such as recommending treatments in clinical settings, is important for enhancing clinical decision-support systems. However, the challenge lies in accurately evaluating and optimizing these policies for maximum efficacy. This paper addresses this gap by focusing on two key aspects of policy learning: evaluation and optimization.

METHOD

We develop counterfactual policy learning algorithms for practical clinical applications to suggest viable treatment for patients. We first design a bootstrap method for counterfactual assessment and enhancement of policies, aiming to diminish uncertainty in clinical decisions. Building on this, we introduce an innovative adversarial learning algorithm, inspired by bootstrap principles, to further advance policy optimization.

RESULTS

The efficacy of our algorithms was validated using both semi-synthetic and real-world clinical datasets. Our method outperforms baseline algorithms, reducing the variance in policy evaluation by 30% and the error rate by 25%. In policy optimization, it enhances the reward by 1% to 3%, highlighting the practical value of our approach in clinical decision-making.

CONCLUSION

This study demonstrates the effectiveness of combining bootstrap and adversarial learning techniques in policy learning for clinical decision support. It not only enhances the accuracy and reliability of policy evaluation and optimization but also paves avenues for leveraging advanced counterfactual machine learning in healthcare.

Effect of Vascular Senescence on the Efficacy and Safety of Warfarin: Insights from Rat Models and a Prospective Cohort Study

Warfarin, with its narrow therapeutic range, requires the understanding of various influencing factors for personalized medication. Vascular senescence, marked by vascular stiffening and endothelial dysfunction, has an unclear effect on the efficacy and safety of warfarin. Based on previous studies, we hypothesized that vascular senescence increases the risk of bleeding during warfarin therapy. This study aimed to explore these effects using animal models and clinical cohorts. We established rat models of vascular senescence and calcification using d-galactose, vitamin D, and nicotine. After validating the models, we examined changes in the international normalized ratio (INR) at fixed warfarin doses (0.20 and 0.35 mg/kg). We found that vascular senescence caused significantly elevated INR values and increased bleeding risk. In the prospective clinical cohort study (NCT06428110), hospitalized warfarin patients with standard dose adjustments were divided into vascular senescence and control groups based on ultrasound and computed tomography diagnosis. Using propensity score matching to exclude the influence of confounding factors, we found that the vascular senescence group had lower steady-state warfarin doses and larger dose adjustments, with a higher probability of INR exceeding the therapeutic range. The vascular senescence group tended to experience more bleeding or thromboembolic/ischemic events during 1 year of follow-up, while there was no statistical difference. In conclusion, vascular senescence leads to unstable INR values and increases higher bleeding risk during warfarin therapy, highlighting the importance of considering vascular senescence in future precision warfarin therapies. SIGNIFICANCE STATEMENT: Many factors influence warfarin efficacy; however, the effect of vascular senescence remains unclear. This study aimed to investigate the effects of vascular senescence on the efficacy and safety of warfarin. Through both rat models and clinical cohort studies, our findings indicated that vascular senescence may compromise the stability of warfarin, presenting challenges in maintaining its efficacy and safety.

Leveraging QSP Models for MIPD: A Case Study for Warfarin/INR

Warfarin dosing remains challenging due to substantial inter-individual variability, which can lead to unsafe or ineffective therapy with standard dosing. Model-informed precision dosing (MIPD) can help individualize warfarin dosing, requiring the selection of a suitable model. For models developed from clinical data, the dependence on the study design and population raises questions about generalizability. Quantitative system pharmacology (QSP) models promise better extrapolation abilities; however, their complexity and lack of validation on clinical data raise questions about applicability in MIPD. We have previously derived a mechanistic warfarin/international normalized ratio (INR) model from a blood coagulation QSP model. In this article, we evaluated the predictive performance of the warfarin/INR model in the context of MIPD using an external dataset with INR data from patients starting warfarin treatment. We assessed the accuracy and precision of model predictions, benchmarked against an empirically based reference model. Additionally, we evaluated covariate contributions and assessed the predictive performance separately in the more challenging outpatient data. The warfarin/INR model performed comparably to the reference model across various measures despite not being calibrated with warfarin initiation data. Including CYP2C9 and/or VKORC1 genotypes as covariates improved the prediction quality of the warfarin/INR model, even after assimilating 4 days of INR data. The outpatient INR exhibited higher unexplained variability, and predictions slightly exceeded observed values, suggesting that model adjustments might be necessary when transitioning from an inpatient to an outpatient setting. Overall, this research underscores the potential of QSP-derived models for MIPD, offering a complementary approach to empirical model development.

Cardiovascular Pharmacogenetics: From Discovery of Genetic Association to Clinical Adoption of Derived Test

Recent breakthroughs in human genetics and in information technologies have markedly expanded our understanding at the molecular level of the response to drugs, i.e., pharmacogenetics (PGx), across therapy areas. This review is restricted to PGx for cardiovascular (CV) drugs. First, we examined the PGx information in the labels approved by regulatory agencies in Europe, Japan, and North America and related recommendations from expert panels. Out of 221 marketed CV drugs, 36 had PGx information in their labels approved by one or more agencies. The level of annotations and recommendations varied markedly between agencies and expert panels. Clopidogrel is the only CV drug with consistent PGx recommendation (i.e., "actionable"). This situation prompted us to dissect the steps from discovery of a PGx association to clinical translation. We found 101 genome-wide association studies that investigated the response to CV drugs or drug classes. These studies reported significant associations for 48 PGx traits mapping to 306 genes. Six of these 306 genes are mentioned in the corresponding PGx labels or recommendations for CV drugs. Genomic analyses also highlighted the wide between-population differences in risk allele frequencies and the individual load of actionable PGx variants. Given the high attrition rate and the long road to clinical translation, additional work is warranted to identify and validate PGx variants for more CV drugs across diverse populations and to demonstrate the utility of PGx testing. To that end, pre-emptive PGx combining genomic profiling with electronic medical records opens unprecedented opportunities to improve healthcare, for CV diseases and beyond. SIGNIFICANCE STATEMENT: Despite spectacular breakthroughs in human molecular genetics and information technologies, consistent evidence supporting PGx testing in the cardiovascular area is limited to a few drugs. Additional work is warranted to discover and validate new PGx markers and demonstrate their utility. Pre-emptive PGx combining genomic profiling with electronic medical records opens unprecedented opportunities to improve healthcare, for CV diseases and beyond.

Implementation of pharmacogenomics: Where are we now?

Pharmacogenomics (PGx), examining the effect of genetic variation on interpatient variation in drug disposition and response, has been widely studied for several decades. However, as cost, as well as turnaround time associated with PGx testing, has significantly improved, the use of PGx in the clinical setting has been gaining momentum. Nevertheless, challenges have emerged in the broader clinical implementation of PGx. In this review, we will outline current models of PGx delivery and methodologies of evaluation, and discuss clinically relevant PGx tests and associated medications. Additionally, we will describe our approach for the broad implementation of pre-emptive DPYD genotyping in patients taking fluoropyrimidines in Ontario, Canada, as an example of clinically actionable PGx testing with sufficient clinical evidence of patient benefit that can become a new standard of patient care. We will highlight challenges associated with PGx testing, including a lack of diversity in PGx studies as well as general limitations that impact the broad adoption of PGx testing. Lastly, we examine the future of PGx, discussing new clinical targets, methodologies and analysis approaches.

Gene-environment interactions within a precision environmental health framework

Understanding the complex interplay of genetic and environmental factors in disease etiology and the role of gene-environment interactions (GEIs) across human development stages is important. We review the state of GEI research, including challenges in measuring environmental factors and advantages of GEI analysis in understanding disease mechanisms. We discuss the evolution of GEI studies from candidate gene-environment studies to genome-wide interaction studies (GWISs) and the role of multi-omics in mediating GEI effects. We review advancements in GEI analysis methods and the importance of large-scale datasets. We also address the translation of GEI findings into precision environmental health (PEH), showcasing real-world applications in healthcare and disease prevention. Additionally, we highlight societal considerations in GEI research, including environmental justice, the return of results to participants, and data privacy. Overall, we underscore the significance of GEI for disease prediction and prevention and advocate for integrating the exposome into PEH omics studies.

Data science through natural language with ChatGPT's Code Interpreter

Large language models (LLMs) have emerged as a powerful tool for biomedical researchers, demonstrating remarkable capabilities in understanding and generating human-like text. ChatGPT with its Code Interpreter functionality, an LLM connected with the ability to write and execute code, streamlines data analysis workflows by enabling natural language interactions. Using materials from a previously published tutorial, similar analyses can be performed through conversational interactions with the chatbot, covering data loading and exploration, model development and comparison, permutation feature importance, partial dependence plots, and additional analyses and recommendations. The findings highlight the significant potential of LLMs in assisting researchers with data analysis tasks, allowing them to focus on higher-level aspects of their work. However, there are limitations and potential concerns associated with the use of LLMs, such as the importance of critical thinking, privacy, security, and equitable access to these tools. As LLMs continue to improve and integrate with available tools, data science may experience a transformation similar to the shift from manual to automatic transmission in driving. The advancements in LLMs call for considering the future directions of data science and its education, ensuring that the benefits of these powerful tools are utilized with proper human supervision and responsibility.

Clinical versus fixed warfarin dosing and the impact on quality of anticoagulation (The ClinFix trial)

Different dosing strategies exist to initiate warfarin, most commonly fixed warfarin dosing (FWD), clinical warfarin dosing (CWD), and genetic-guided warfarin dosing (GWD). Landmark trials have shown GWD to be superior when compared to FWD in the EU-PACT trial or CWD in the GIFT trial. COAG trial did not show differences between GWD and CWD. We aim to compare the anticoagulation quality outcomes of CWD and FWD. This is a prospective cohort study with a retrospective comparator. Recruited subjects in the CWD (prospective) arm were initiated on warfarin according to the clinical dosing component of the algorithm published in www.warfarindosing.org. The primary efficacy outcome was the percentage time in the therapeutic range (PTTR) from day 3 to 6 till day 28 to 35. The study enrolled 122 and 123 patients in the CWD and FWD, respectively. The PTTR did not differ statistically between CWD and FWD (62.2 ± 26.2% vs. 58 ± 25.4%, p = 0.2). There was also no difference between both arms in the percentage of visits with extreme subtherapeutic international normalized ratio (INR) (<1.5; 15 ± 18.3% vs. 16.8 ± 19.1%, p = 0.44) or extreme supratherapeutic INR (>4; 7.7 ± 14.7% vs. 7.5 ± 12.4%, p = 0.92). We conclude that CWD did not improve the anticoagulation quality parameters compared to the FWD method.

QPGx-CARES: Qatar pharmacogenetics clinical applications and research enhancement strategies

Pharmacogenetic (PGx)-informed medication prescription is a cutting-edge genomic application in contemporary medicine, offering the potential to overcome the conventional "trial-and-error" approach in drug prescription. The ability to use an individual's genetic profile to predict drug responses allows for personalized drug and dosage selection, thereby enhancing the safety and efficacy of treatments. However, despite significant scientific and clinical advancements in PGx, its integration into routine healthcare practices remains limited. To address this gap, the Qatar Genome Program (QGP) has embarked on an ambitious initiative known as QPGx-CARES (Qatar Pharmacogenetics Clinical Applications and Research Enhancement Strategies), which aims to set a roadmap for optimizing PGx research and clinical implementation on a national scale. The goal of QPGx-CARES initiative is to integrate PGx testing into clinical settings with the aim of improving patient health outcomes. In 2022, QGP initiated several implementation projects in various clinical settings. These projects aimed to evaluate the clinical utility of PGx testing, gather valuable insights into the effective dissemination of PGx data to healthcare professionals and patients, and identify the gaps and the challenges for wider adoption. QPGx-CARES strategy aimed to integrate evidence-based PGx findings into clinical practice, focusing on implementing PGx testing for cardiovascular medications, supported by robust scientific evidence. The current initiative sets a precedent for the nationwide implementation of precision medicine across diverse clinical domains.

Warfarin-A natural anticoagulant: A review of research trends for precision medication

BACKGROUND

Warfarin is a widely prescribed anticoagulant in the clinic. It has a more considerable individual variability, and many factors affect its variability. Mathematical models can quantify the quantitative impact of these factors on individual variability.

PURPOSE

The aim is to comprehensively analyze the advanced warfarin dosing algorithm based on pharmacometrics and machine learning models of personalized warfarin dosage.

METHODS

A bibliometric analysis of the literature retrieved from PubMed and Scopus was performed using VOSviewer. The relevant literature that reported the precise dosage of warfarin calculation was retrieved from the database. The multiple linear regression (MLR) algorithm was excluded because a recent systematic review that mainly reviewed this algorithm has been reported. The following terms of quantitative systems pharmacology, mechanistic model, physiologically based pharmacokinetic model, artificial intelligence, machine learning, pharmacokinetic, pharmacodynamic, pharmacokinetics, pharmacodynamics, and warfarin were added as MeSH Terms or appearing in Title/Abstract into query box of PubMed, then humans and English as filter were added to retrieve the literature.

RESULTS

Bibliometric analysis revealed important co-occuring MeShH and index keywords. Further, the United States, China, and the United Kingdom were among the top countries contributing in this domain. Some studies have established personalized warfarin dosage models using pharmacometrics and machine learning-based algorithms. There were 54 related studies, including 14 pharmacometric models, 31 artificial intelligence models, and 9 model evaluations. Each model has its advantages and disadvantages. The pharmacometric model contains biological or pharmacological mechanisms in structure. The process of pharmacometric model development is very time- and labor-intensive. Machine learning is a purely data-driven approach; its parameters are more mathematical and have less biological interpretation. However, it is faster, more efficient, and less time-consuming. Most published models of machine learning algorithms were established based on cross-sectional data sourced from the database.

CONCLUSION

Future research on personalized warfarin medication should focus on combining the advantages of machine learning and pharmacometrics algorithms to establish a more robust warfarin dosage algorithm. Randomized controlled trials should be performed to evaluate the established algorithm of warfarin dosage. Moreover, a more user-friendly and accessible warfarin precision medicine platform should be developed.

Stability of Anticoagulation Following Acenocoumarin in Stroke Patients: Role of Pharmacogenomics and Acquired Factors

OBJECTIVE

Pharmacogenomics plays an important role in drug metabolism. A stable anticoagulation is important for primary and secondary prevention of cardioembolic stroke and cerebral venous sinus thrombosis (CVST). We report the role of cytochrome P450 ( CYP2C9*2/*3 ) and vitamin K epoxide reductase subunit 1 ( VKORC1 ) genotypes and acquired causes in maintaining stability of anticoagulation following acenocoumarin in cardioembolic stroke and CVST.

METHODS

The study comprised 157 individuals with cardioembolic stroke and CVST who were on acenocoumarin. Their comorbidities, comedication, and dietary habits were noted. Prothrombin time and international normalized ratio (INR) were measured during follow-up, and the coagulation status was categorized as stable (>50% occasions in therapeutic range) and unstable (>50% below and above therapeutic range). Genotyping of VKORC1 , CYP2C9*2 , and CYP2C9*3 was done by polymerase chain reaction-restriction fragment length polymorphism. Bleeding and embolic complications were noted. The predictors of unstable INR were evaluated using multivariate analysis.

RESULTS

INR was stable in 47.8% and unstable in 52.2% of patients. Patients with mutant genotypes required low dose of acenocoumarin. The predictors of unstable INR were metallic valve (odds ratio [OR] 4.07, 95% confidence interval [CI] 1.23-13.49, P = 0.02), use of digoxin (OR 0.031, 95% CI 0.13-0.74, P = 0.09), proton pump inhibitor (OR 0.23, 95% CI 0.06-0.91, P = 0.037), sodium valproate (OR 0.22, 95% CI 0.05-0.85, P = 0.029), and CYP2C9*2 genotype (OR 5.57, 95% CI 1.19-26.06, P = 0.02).

CONCLUSIONS

Variant genotypes of VKORC1 , CYP2C9*2 , and CYP2C9*3 required lower dose of acenocoumarin, and CYP2C9*2 was associated with unstable INR. Comedication is a modifiable risk factor that needs attention.

Pregnancy outcomes in Chinese women with mechanical heart valves receiving warfarin treatment throughout pregnancy: 14-year experience

OBJECTIVE

The purpose was to evaluate pregnancy outcomes and risk factors associated with fetal complications in Chinese pregnant women with mechanical heart valves (MHVs) taking low-dose warfarin, aiming to fill in the research gap of this area.

METHODS

Between June 2010 and Aug 2023, 122 patients with MHVs who had 151 pregnancies and received warfarin throughout pregnancy were included. We compared them with 302 paired pregnancies without warfarin treatment. Binary logistic regression analyses were performed to explore risk predictors of fetal complications.

RESULTS

Pregnancy loss rate was 37.1 % in women taking warfarin, compared to only 4.6 % for those without warfarin exposure in pregnancy (RR = 8.00, 95 % CI: 4.61-13.90). In pregnant women with MHVs, there were 34 spontaneous abortions, 22 stillbirths and 1 neonatal malformation. In the first, second and third pregnant trimesters of women with MHVs, fetal complication incidences were 19.2 %, 9.9 % and 8.0 %, respectively. 86.0 % of fetal complications occurred in women taking a warfarin dose ≤5 mg/d, accounting for 94.0 % of the total population. The newborns' birth weight, gestational age and 1-minute Apgar score were significantly lower in pregnancies treated with warfarin compared to those without warfarin exposure. Only 2.0 % of postpartum hemorrhage and no thrombosis or maternal mortality data were collected in pregnant women on warfarin in this study.

CONCLUSION

Most Chinese pregnant women take a warfarin daily dose ≤5 mg and they might have only around 60 % chance of giving birth to a live baby without maternal complications.

Being precise with anticoagulation to reduce adverse drug reactions: are we there yet?

Anticoagulants are potent therapeutics widely used in medical and surgical settings, and the amount spent on anticoagulation is rising. Although warfarin remains a widely prescribed oral anticoagulant, prescriptions of direct oral anticoagulants (DOACs) have increased rapidly. Heparin-based parenteral anticoagulants include both unfractionated and low molecular weight heparins (LMWHs). In clinical practice, anticoagulants are generally well tolerated, although interindividual variability in response is apparent. This variability in anticoagulant response can lead to serious incident thrombosis, haemorrhage and off-target adverse reactions such as heparin-induced thrombocytopaenia (HIT). This review seeks to highlight the genetic, environmental and clinical factors associated with variability in anticoagulant response, and review the current evidence base for tailoring the drug, dose, and/or monitoring decisions to identified patient subgroups to improve anticoagulant safety. Areas that would benefit from further research are also identified. Validated variants in VKORC1, CYP2C9 and CYP4F2 constitute biomarkers for differential warfarin response and genotype-informed warfarin dosing has been shown to reduce adverse clinical events. Polymorphisms in CES1 appear relevant to dabigatran exposure but the genetic studies focusing on clinical outcomes such as bleeding are sparse. The influence of body weight on LMWH response merits further attention, as does the relationship between anti-Xa levels and clinical outcomes. Ultimately, safe and effective anticoagulation requires both a deeper parsing of factors contributing to variable response, and further prospective studies to determine optimal therapeutic strategies in identified higher risk subgroups.

Optimizing warfarin dosing for patients with atrial fibrillation using machine learning

While novel oral anticoagulants are increasingly used to reduce risk of stroke in patients with atrial fibrillation, vitamin K antagonists such as warfarin continue to be used extensively for stroke prevention across the world. While effective in reducing the risk of strokes, the complex pharmacodynamics of warfarin make it difficult to use clinically, with many patients experiencing under- and/or over- anticoagulation. In this study we employed a novel implementation of deep reinforcement learning to provide clinical decision support to optimize time in therapeutic International Normalized Ratio (INR) range. We used a novel semi-Markov decision process formulation of the Batch-Constrained deep Q-learning algorithm to develop a reinforcement learning model to dynamically recommend optimal warfarin dosing to achieve INR of 2.0-3.0 for patients with atrial fibrillation. The model was developed using data from 22,502 patients in the warfarin treated groups of the pivotal randomized clinical trials of edoxaban (ENGAGE AF-TIMI 48), apixaban (ARISTOTLE) and rivaroxaban (ROCKET AF). The model was externally validated on data from 5730 warfarin-treated patients in a fourth trial of dabigatran (RE-LY) using multilevel regression models to estimate the relationship between center-level algorithm consistent dosing, time in therapeutic INR range (TTR), and a composite clinical outcome of stroke, systemic embolism or major hemorrhage. External validation showed a positive association between center-level algorithm-consistent dosing and TTR (R2 = 0.56). Each 10% increase in algorithm-consistent dosing at the center level independently predicted a 6.78% improvement in TTR (95% CI 6.29, 7.28; p < 0.001) and a 11% decrease in the composite clinical outcome (HR 0.89; 95% CI 0.81, 1.00; p = 0.015). These results were comparable to those of a rules-based clinical algorithm used for benchmarking, for which each 10% increase in algorithm-consistent dosing independently predicted a 6.10% increase in TTR (95% CI 5.67, 6.54, p < 0.001) and a 10% decrease in the composite outcome (HR 0.90; 95% CI 0.83, 0.98, p = 0.018). Our findings suggest that a deep reinforcement learning algorithm can optimize time in therapeutic range for patients taking warfarin. A digital clinical decision support system to promote algorithm-consistent warfarin dosing could optimize time in therapeutic range and improve clinical outcomes in atrial fibrillation globally.

CAGI, the Critical Assessment of Genome Interpretation, establishes progress and prospects for computational genetic variant interpretation methods

BACKGROUND

The Critical Assessment of Genome Interpretation (CAGI) aims to advance the state-of-the-art for computational prediction of genetic variant impact, particularly where relevant to disease. The five complete editions of the CAGI community experiment comprised 50 challenges, in which participants made blind predictions of phenotypes from genetic data, and these were evaluated by independent assessors.

RESULTS

Performance was particularly strong for clinical pathogenic variants, including some difficult-to-diagnose cases, and extends to interpretation of cancer-related variants. Missense variant interpretation methods were able to estimate biochemical effects with increasing accuracy. Assessment of methods for regulatory variants and complex trait disease risk was less definitive and indicates performance potentially suitable for auxiliary use in the clinic.

CONCLUSIONS

Results show that while current methods are imperfect, they have major utility for research and clinical applications. Emerging methods and increasingly large, robust datasets for training and assessment promise further progress ahead.

Improved stacking ensemble learning based on feature selection to accurately predict warfarin dose

Background

With the rapid development of artificial intelligence, prediction of warfarin dose via machine learning has received more and more attention. Since the dose prediction involve both linear and nonlinear problems, traditional machine learning algorithms are ineffective to solve such problems at one time.

Objective

Based on the characteristics of clinical data of Chinese warfarin patients, an improved stacking ensemble learning can achieve higher prediction accuracy.

Methods

Information of 641 patients from southern China who had reached a steady state on warfarin was collected, including demographic information, medical history, genotype, and co-medication status. The dataset was randomly divided into a training set (90%) and a test set (10%). The predictive capability is evaluated on a new test set generated by stacking ensemble learning. Additional factors associated with warfarin dose were discovered by feature selection methods.

Results

A newly proposed heuristic-stacking ensemble learning performs better than traditional-stacking ensemble learning in key metrics such as accuracy of ideal dose (73.44%, 71.88%), mean absolute errors (0.11 mg/day, 0.13 mg/day), root mean square errors (0.18 mg/day, 0.20 mg/day) and R2 (0.87, 0.82).

Conclusions

The developed heuristic-stacking ensemble learning can satisfactorily predict warfarin dose with high accuracy. A relationship between hypertension, a history of severe preoperative embolism, and warfarin dose is found, which provides a useful reference for the warfarin dose administration in the future.

New insights into the role of VKORC1 polymorphisms for optimal warfarin dose selection in Caribbean Hispanic patients through an external validation of a population PK/PD model

Warfarin, an oral anticoagulant, has been used for decades to prevent thromboembolic events. The complex interplay between CYP2C9 and VKORC1 genotypes on warfarin PK and PD properties is not fully understood in special sub-groups of patients. This study aimed to externally validate a population pharmacokinetic/pharmacodynamic (PK/PD) model for the effect of warfarin on international normalized ratio (INR) and to evaluate optimal dosing strategies based on the selected covariates in Caribbean Hispanic patients. INR, and CYP2C9 and VKORC1 genotypes from 138 patients were used to develop a population PK/PD model in NONMEM. The structural definition of a previously published PD model for INR was implemented. A numerical evaluation of the parameter-covariate relationship was performed. Simulations were conducted to determine optimal dosing strategies for each genotype combinations, focusing on achieving therapeutic INR levels. Findings revealed elevated IC50 for G/G, G/A, and A/A VKORC1 haplotypes (11.76, 10.49, and 9.22 mg/L, respectively), in this population compared to previous reports. The model-guided dosing analysis recommended daily warfarin doses of 3-5 mg for most genotypes to maintain desired INR levels, although subjects with combination of CYP2C9 and VKORC1 genotypes * 2/* 2-, * 2/* 3- and * 2/* 5-A/A would require only 1 mg daily. This research underscores the potential of population PK/PD modeling to inform personalized warfarin dosing in populations typically underrepresented in clinical studies, potentially leading to improved treatment outcomes and patient safety. By integrating genetic factors and clinical data, this approach could pave the way for more effective and tailored anticoagulation therapy in diverse patient groups.

Evaluation of Machine Learning Approaches for Predicting Warfarin Discharge Dose in Cardiac Surgery Patients: Retrospective Algorithm Development and Validation Study

BACKGROUND

Warfarin dosing in cardiac surgery patients is complicated by a heightened sensitivity to the drug, predisposing patients to adverse events. Predictive algorithms are therefore needed to guide warfarin dosing in cardiac surgery patients.

OBJECTIVE

This study aimed to develop and validate an algorithm for predicting the warfarin dose needed to attain a therapeutic international normalized ratio (INR) at the time of discharge in cardiac surgery patients.

METHODS

We abstracted variables influencing warfarin dosage from the records of 1031 encounters initiating warfarin between April 1, 2011, and November 29, 2019, at St Michael's Hospital in Toronto, Ontario, Canada. We compared the performance of penalized linear regression, k-nearest neighbors, random forest regression, gradient boosting, multivariate adaptive regression splines, and an ensemble model combining the predictions of the 5 regression models. We developed and validated separate models for predicting the warfarin dose required for achieving a discharge INR of 2.0-3.0 in patients undergoing all forms of cardiac surgery except mechanical mitral valve replacement and a discharge INR of 2.5-3.5 in patients receiving a mechanical mitral valve replacement. For the former, we selected 80% of encounters (n=780) who had initiated warfarin during their hospital admission and had achieved a target INR of 2.0-3.0 at the time of discharge as the training cohort. Following 10-fold cross-validation, model accuracy was evaluated in a test cohort comprised solely of cardiac surgery patients. For patients requiring a target INR of 2.5-3.5 (n=165), we used leave-p-out cross-validation (p=3 observations) to estimate model performance. For each approach, we determined the mean absolute error (MAE) and the proportion of predictions within 20% of the true warfarin dose. We retrospectively evaluated the best-performing algorithm in clinical practice by comparing the proportion of cardiovascular surgery patients discharged with a therapeutic INR before (April 2011 and July 2019) and following (September 2021 and May 2, 2022) its implementation in routine care.

RESULTS

Random forest regression was the best-performing model for patients with a target INR of 2.0-3.0, an MAE of 1.13 mg, and 39.5% of predictions of falling within 20% of the actual therapeutic discharge dose. For patients with a target INR of 2.5-3.5, the ensemble model performed best, with an MAE of 1.11 mg and 43.6% of predictions being within 20% of the actual therapeutic discharge dose. The proportion of cardiovascular surgery patients discharged with a therapeutic INR before and following implementation of these algorithms in clinical practice was 47.5% (305/641) and 61.1% (11/18), respectively.

CONCLUSIONS

Machine learning algorithms based on routinely available clinical data can help guide initial warfarin dosing in cardiac surgery patients and optimize the postsurgical anticoagulation of these patients.

Exploring the complex relationship between vitamin K, gut microbiota, and warfarin variability in cardiac surgery patients

BACKGROUND AND OBJECTIVES

Due to the high individual variability of anticoagulant warfarin, this study aimed to investigate the effects of vitamin K concentration and gut microbiota on individual variability of warfarin in 246 cardiac surgery patients.

METHODS

The pharmacokinetics and pharmacodynamics (PKPD) model predicted international normalized ratio (INR) and warfarin concentration. Serum and fecal samples were collected to detect warfarin and vitamin K [VK1 and menaquinone-4 (MK4)] concentrations and gut microbiota diversity, respectively. In addition, the patient's medical records were reviewed for demographic characteristics, drug history, and CYP2C9, VKORC1, and CYP4F2 genotypes.

RESULTS

The PKPD model predicted ideal values of 62.7% for S-warfarin, 70.4% for R-warfarin, and 76.4% for INR. The normal VK1 level was 1.34±1.12 nmol/ml (95% CI: 0.33-4.08 nmol/ml), and the normal MK4 level was 0.22±0.18 nmol/ml (95% CI: 0.07-0.63 nmol/ml). The MK4 to total vitamin K ratio was 16.5±9.8% (95% CI: 4.3-41.5%). The S-warfarin concentration of producing 50% of maximum anticoagulation and the half-life of prothrombin complex activity tended to increase with vitamin K. Further, Prevotella and Eubacterium of gut microbiota identified as the main bacteria associated with individual variability of warfarin. The results suggest that an increase in vitamin K concentration can decrease anticoagulation, and gut microbiota may influence warfarin anticoagulation through vitamin K2 synthesis.

CONCLUSION

This study highlights the importance of considering vitamin K concentration and gut microbiota when prescribing warfarin. The findings may have significant implications for the personalized use of warfarin. Further research is needed to understand better the role of vitamin K and gut microbiota in warfarin anticoagulation.

Case studies of clinical decision-making through prescriptive models based on machine learning

BACKGROUND

The development of computational methodologies to support clinical decision-making is of vital importance to reduce morbidity and mortality rates. Specifically, prescriptive analytic is a promising area to support decision-making in the monitoring, treatment and prevention of diseases. These aspects remain a challenge for medical professionals and health authorities.

MATERIALS AND METHODS

In this study, we propose a methodology for the development of prescriptive models to support decision-making in clinical settings. The prescriptive model requires a predictive model to build the prescriptions. The predictive model is developed using fuzzy cognitive maps and the particle swarm optimization algorithm, while the prescriptive model is developed with an extension of fuzzy cognitive maps that combines them with genetic algorithms. We evaluated the proposed approach in three case studies related to monitoring (warfarin dose estimation), treatment (severe dengue) and prevention (geohelminthiasis) of diseases.

RESULTS

The performance of the developed prescriptive models demonstrated the ability to estimate warfarin doses in coagulated patients, prescribe treatment for severe dengue and generate actions aimed at the prevention of geohelminthiasis. Additionally, the predictive models can predict coagulation indices, severe dengue mortality and soil-transmitted helminth infections.

CONCLUSIONS

The developed models performed well to prescribe actions aimed to monitor, treat and prevent diseases. This type of strategy allows supporting decision-making in clinical settings. However, validations in health institutions are required for their implementation.

Machine learning in medication prescription: A systematic review

BACKGROUND

Medication prescription is a complex process that could benefit from current research and development in machine learning through decision support systems. Particularly pediatricians are forced to prescribe medications "off-label" as children are still underrepresented in clinical studies, which leads to a high risk of an incorrect dose and adverse drug effects.

METHODS

PubMed, IEEE Xplore and PROSPERO were searched for relevant studies that developed and evaluated well-performing machine learning algorithms following the PRISMA statement. Quality assessment was conducted in accordance with the IJMEDI checklist. Identified studies were reviewed in detail, including the required variables for predicting the correct dose, especially of pediatric medication prescription.

RESULTS

The search identified 656 studies, of which 64 were reviewed in detail and 36 met the inclusion criteria. According to the IJMEDI checklist, five studies were considered to be of high quality. 19 of the 36 studies dealt with the active substance warfarin. Overall, machine learning algorithms based on decision trees or regression methods performed superior regarding their predictive power than algorithms based on neural networks, support vector machines or other methods. The use of ensemble methods like bagging or boosting generally enhanced the accuracy of the dose predictions. The required input and output variables of the algorithms were considerably heterogeneous and differ strongly among the respective substance.

CONCLUSIONS

By using machine learning algorithms, the prescription process could be simplified and dosing correctness could be enhanced. Despite the heterogenous results among the different substances and cases and the lack of pediatric use cases, the identified approaches and required variables can serve as an excellent starting point for further development of algorithms predicting drug doses, particularly for children. Especially the combination of physiologically-based pharmacokinetic models with machine learning algorithms represents a great opportunity to enhance the predictive power and accuracy of the developed algorithms.

Effect of MEF2A and SLC22A3-LPAL2-LPA gene polymorphisms on warfarin sensitivity and responsiveness in Jordanian cardiovascular patients

AIMS

This study aims to investigate the influence of MEF2A and SLC22A3-LPAL2-LPA polymorphisms on cardiovascular disease susceptibility and responsiveness to warfarin medication in Jordanian patients, during the initiation and maintenance phases of treatment.

BACKGROUNDS

Several candidate genes have been reported to be involved in warfarin metabolism and studying such genes may help in finding an accurate way to determine the needed warfarin dose to lower the risk of adverse drug effects, resulting in more safe anticoagulant therapy.

METHODS

The study population included 212 cardiovascular patients and 213 healthy controls. Genotyping of MEF2A and SLC22A3-LPAL2-LPA polymorphisms was conducted to examine their effects on warfarin efficiency and cardiovascular disease susceptibility using PCR-based methods.

RESULTS

One SNP (SLC22A3-LPAL2-LPA rs10455872) has been associated with cardiovascular disease in the Jordanian population, whereas the other SNPs in the MEF2A gene and SLC22A3-LPAL2-LPA gene cluster did not have any significant differences between cardiovascular patients and healthy individuals. Moreover, SLC22A3-LPAL2-LPA rs10455872 was correlated with moderate warfarin sensitivity, the other SNPs examined in the current study have not shown any significant associations with warfarin sensitivity and responsiveness.

CONCLUSION

Our data refer to a lack of correlation between the MEF2A polymorphism and the efficacy of warfarin treatment in both phases of treatment, the initiation, and maintenance phases. However, only rs10455872 SNP was associated with sensitivity to warfarin during the initiation phase. Furthermore, rs3125050 has been found to be associated with the international normalized number treatment outcomes in the maintenance phase.

Simulating clinical trials for model-informed precision dosing: using warfarin treatment as a use case

Treatment response variability across patients is a common phenomenon in clinical practice. For many drugs this inter-individual variability does not require much (if any) individualisation of dosing strategies. However, for some drugs, including chemotherapies and some monoclonal antibody treatments, individualisation of dosages are needed to avoid harmful adverse events. Model-informed precision dosing (MIPD) is an emerging approach to guide the individualisation of dosing regimens of otherwise difficult-to-administer drugs. Several MIPD approaches have been suggested to predict dosing strategies, including regression, reinforcement learning (RL) and pharmacokinetic and pharmacodynamic (PKPD) modelling. A unified framework to study the strengths and limitations of these approaches is missing. We develop a framework to simulate clinical MIPD trials, providing a cost and time efficient way to test different MIPD approaches. Central for our framework is a clinical trial model that emulates the complexities in clinical practice that challenge successful treatment individualisation. We demonstrate this framework using warfarin treatment as a use case and investigate three popular MIPD methods: 1. Neural network regression; 2. Deep RL; and 3. PKPD modelling. We find that the PKPD model individualises warfarin dosing regimens with the highest success rate and the highest efficiency: 75.1% of the individuals display INRs inside the therapeutic range at the end of the simulated trial; and the median time in the therapeutic range (TTR) is 74%. In comparison, the regression model and the deep RL model have success rates of 47.0% and 65.8%, and median TTRs of 45% and 68%. We also find that the MIPD models can attain different degrees of individualisation: the Regression model individualises dosing regimens up to variability explained by covariates; the Deep RL model and the PKPD model individualise dosing regimens accounting also for additional variation using monitoring data. However, the Deep RL model focusses on control of the treatment response, while the PKPD model uses the data also to further the individualisation of predictions.

Optimisation of warfarin-dosing algorithms for Han Chinese patients with CYP2C9*13 variants

BACKGROUND

Existing pharmacogenetic algorithms cannot fully explain warfarin dose variability in all patients. CYP2C9*13 is an important allelic variant in the Han Chinese population. However, adjustment of warfarin dosing in CYP2C9*13 variant carriers remains unclear. To the best of our knowledge, this study is the first to assess the effects of adjusting warfarin dosages in Han Chinese patients harbouring CYP2C9*13 variants.

METHODS

In total, 971 warfarin-treated Han Chinese patients with atrial fibrillation were enrolled in this study. Clinical data were collected, and CYP2C9*2, *3, *13 and VKORC1-1639 G > A variants were genotyped. We quantitatively analysed the effect of CYP2C9*13 on warfarin maintenance dose and provided multiplicative adjustments for CYP2C9*13 using validated pharmacogenetic algorithms.

RESULTS

Approximately 0.6% of the Han Chinese population carried CYP2C9*13 variant, and the genotype frequency was between those of CYP2C9*2 and CYP2C9*3. The warfarin maintenance doses were significantly reduced in CYP2C9*13 carriers. When CYP2C9*13 variants were not considered, the pharmacogenetic algorithms overestimated warfarin maintenance doses by 1.03-1.16 mg/d on average. The actual warfarin dose in CYP2C9*13 variant carriers was approximately 40% lower than the algorithm-predicted dose. Adjusting the warfarin-dosing algorithm according to the CYP2C9*13 allele could reduce the dose prediction error.

CONCLUSION

Our study showed that the algorithm-predicted doses should be lowered for CYP2C9*13 carriers. Inclusion of the CYP2C9*13 variant in the warfarin-dosing algorithm tends to predict the warfarin maintenance dose more accurately and improves the efficacy and safety of warfarin administration in Han Chinese patients.

An Adaptive Model Predictive Controller to Address the Biovariability in Blood Clotting Response During Therapy With Warfarin

OBJECTIVE

Effective dosing of anticoagulants aims to prevent blood clot formation while avoiding hemorrhages. This complex task is challenged by several disturbing factors and drug-effect uncertainties, requesting frequent monitoring and adjustment. Biovariability in drug absorption and action further complicates titration and calls for individualized strategies. In this paper, we propose an adaptive closed-loop control algorithm to assist in warfarin therapy management.

METHODS

The controller was designed and tested in silico using an established pharmacometrics model of warfarin, which accounts for inter-subject variability. The control algorithm is an adaptive Model Predictive Control (a-MPC) that leverages a simplified patient model, whose parameters are updated with a Bayesian strategy. Performance was quantitatively evaluated in simulations performed on a population of virtual subjects against an algorithm reproducing medical guidelines (MG) and an MPC controller available in the literature (l-MPC).

RESULTS

The proposed a-MPC significantly (p 0.05) lowers rising time (2.8 vs. 4.4 and 11.2 days) and time out of range (3.3 vs. 7.2 and 12.9 days) with respect to both MG and l-MPC, respectively. Adaptivity grants a significantly (p 0.05) lower number of subjects reaching unsafe INR values compared to when this feature is not present (8.9% vs.15% of subjects presenting an overshoot outside the target range and 0.08% vs. 0.28% of subjects reaching dangerous INR values).

CONCLUSION

The a-MPC algorithm improve warfarin therapy compared to the benchmark therapies.

SIGNIFICANCE

This in-silico validation proves effectiveness of the a-MPC algorithm for anticoagulant administration, paving the way for clinical testing.

Development and validation wise assessment of genotype guided warfarin dosing algorithm in Indian population

OBJECTIVES

A study was conducted to develop and validate the warfarin pharmacogenetic dose optimization algorithm considering the clinical pharmacogenetic implementation consortium (CPIC) recommendations for the Asian ethnicity population.

METHODS

The present prospective observational study recruited warfarin-receiving patients. We collected a three ml blood sample for VKORC1, CYP2C9*2, CYP2C9*3, and CYP4F2 polymorphism assessment during the follow-up visits. Clinical history, sociodemographic and warfarin dose details were noted.

RESULTS

The study recruited 300 patients (250 in derivation and 50 in validation timed cohort) receiving warfarin therapy. The baseline characteristics were similar in both cohorts. BMI, presence of comorbidity, VKORC1, CYP2C9*2, and CYP2C9*3 were identified as covariates significantly affecting the warfarin weekly maintenance dose (p<0.001 for all) and the same were included in warfarin pharmacogenetic dose optimization algorithm building. The algorithm built-in the present study showed a good correlation with Gage (r=0.57, p<0.0001), and IWPC (r=0.51, p<0.0001) algorithms, widely accepted in western side of the globe. The receiver operating characteristic curve analysis showed a sensitivity of 73 %, a positive predictive value of 96 %, and a specificity of 89 %. The algorithm correctly identified the validation cohort's warfarin-sensitive, intermediate reacting, and resistant patient populations.

CONCLUSIONS

Validation and comparisons of the warfarin pharmacogenetic dose optimization algorithm have made it ready for the clinical trial assessment.

A "Bundle of Care" to Improve Anticoagulation Control in Patients Receiving Warfarin in Uganda and South Africa: Protocol for an Implementation Study

BACKGROUND

The quality of warfarin anticoagulation among Sub-Saharan African patients is suboptimal. This is due to several factors, including a lack of standardized dosing algorithms, difficulty in providing timely international normalized ratio (INR) results, a lack of patient feedback on their experiences with treatment, a lack of education on adherence, and inadequate knowledge and training of health care workers. Low quality of warfarin anticoagulation, expressed as time in therapeutic range (TTR), is associated with higher adverse event rates, including bleeding and thrombosis, and ultimately, increased morbidity and mortality. Processes and interventions that improve this situation are urgently needed.

OBJECTIVE

This study aims to evaluate the implementation of the "warfarin bundle," a package of interventions to improve the quality of anticoagulation and thereby clinical outcomes. The primary outcome for this study is TTR over the initial 3 months of warfarin therapy.

METHODS

Patients aged 18 years or older who are newly initiated on warfarin for venous thromboembolism, atrial fibrillation, or valvular heart disease will be enrolled and followed up for 3 months at clinics in Cape Town, South Africa, and Kampala, Uganda, where the warfarin bundle is implemented. A retrospective review of the clinical records of patients on warfarin treatment before implementation (controls) will be used for comparison. This study uses a mixed methods approach of the implementation of patient- and process-centered activities to improve the quality of anticoagulation. Patient-centered activities include the use of clinical dosing algorithms, adherence support, and root cause analysis, whereas process-centered activities include point-of-care INR testing, staff training, and patient education and training. We will assess the impact of these interventions by comparing the TTR and safety outcomes across the 2 groups, as well as the cost-effectiveness and acceptability of the package.

RESULTS

We started recruitment in June 2021 and stopped in August 2022, having recruited 167 participants. We obtained ethics approval from the University of Cape Town Faculty of Health Sciences Human Research Ethics Committee, the Provincial Health Research Committees in South Africa, the Joint Clinical Research Centre Institutional Review Board, Kampala, and the University of Liverpool Research Ethics Committee. As of February 2023, data cleaning and formal analysis are underway. We expect to publish the full results by December 2023.

CONCLUSIONS

We anticipate that the "bundle of care," which includes a clinical algorithm to guide individualized dosing of warfarin, will improve INR control and TTR of patients in Uganda and South Africa. We will use these findings to design a larger, multisite clinical trial across several Sub-Saharan African countries.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

DERR1-10.2196/46710.

Cardiovascular precision medicine - A pharmacogenomic perspective

Precision medicine envisages the integration of an individual's clinical and biological features obtained from laboratory tests, imaging, high-throughput omics and health records, to drive a personalised approach to diagnosis and treatment with a higher chance of success. As only up to half of patients respond to medication prescribed following the current one-size-fits-all treatment strategy, the need for a more personalised approach is evident. One of the routes to transforming healthcare through precision medicine is pharmacogenomics (PGx). Around 95% of the population is estimated to carry one or more actionable pharmacogenetic variants and over 75% of adults over 50 years old are on a prescription with a known PGx association. Whilst there are compelling examples of pharmacogenomic implementation in clinical practice, the case for cardiovascular PGx is still evolving. In this review, we shall summarise the current status of PGx in cardiovascular diseases and look at the key enablers and barriers to PGx implementation in clinical practice.

One Rare Warfarin Resistance Case and Possible Mechanism Exploration

One 59-year-old female patient with deep venous thrombosis (DVT) and pulmonary embolism (PE) was treated with 6 mg warfarin once daily as an anticoagulant. Before taking warfarin, her international normalized ratio (INR) was 0.98. Two days after warfarin treatment, her INR did not change from baseline. Due to the high severity of the PE, the patient needed to reach her target range (INR goal = 2.5, range = 2~3) rapidly, so the dose of warfarin was increased from 6 mg daily to 27 mg daily. However, the patient's INR did not improve with the dose escalation, still maintaining an INR of 0.97-0.98. We drew a blood sample half an hour before administering 27 mg warfarin and detected single nucleotide polymorphism for the following genes, which were identified to be relevant with warfarin resistance: CYP2C9 rs1799853, rs1057910, VKORC1 rs9923231, rs61742245, rs7200749, rs55894764, CYP4F2 rs2108622, and GGCX rs2592551. The trough plasma concentration of warfarin was 196.2 ng/mL after 2 days of warfarin administration with 27 mg QD, which was much lower than the therapeutic drug concentration ranges of warfarin (500-3,000 ng/mL). The genotype results demonstrate that the CYP4F2gene has rs2108622 mutation which can explain some aspect of warfarin resistance. Further investigations are necessary to fully characterize other pharmacogenomics or pharmacodynamics determinants of warfarin dose-response in Chinese.

Algorithmic fairness in artificial intelligence for medicine and healthcare

In healthcare, the development and deployment of insufficiently fair systems of artificial intelligence (AI) can undermine the delivery of equitable care. Assessments of AI models stratified across subpopulations have revealed inequalities in how patients are diagnosed, treated and billed. In this Perspective, we outline fairness in machine learning through the lens of healthcare, and discuss how algorithmic biases (in data acquisition, genetic variation and intra-observer labelling variability, in particular) arise in clinical workflows and the resulting healthcare disparities. We also review emerging technology for mitigating biases via disentanglement, federated learning and model explainability, and their role in the development of AI-based software as a medical device.

Missense VKOR mutants exhibit severe warfarin resistance but lack VKCFD via shifting to an aberrantly reduced state

Missense vitamin K epoxide reductase (VKOR) mutations in patients cause resistance to warfarin treatment but not abnormal bleeding due to defective VKOR activity. The underlying mechanism of these phenotypes remains unknown. Here we show that the redox state of these mutants is essential to their activity and warfarin resistance. Using a mass spectrometry-based footprinting method, we found that severe warfarin-resistant mutations change the VKOR active site to an aberrantly reduced state in cells. Molecular dynamics simulation based on our recent crystal structures of VKOR reveals that these mutations induce an artificial opening of the protein conformation that increases access of small molecules, enabling them to reduce the active site and generating constitutive activity uninhibited by warfarin. Increased activity also compensates for the weakened substrate binding caused by these mutations, thereby maintaining normal VKOR function. The uninhibited nature of severe resistance mutations suggests that patients showing signs of such mutations should be treated by alternative anticoagulation strategies.

Approximating Functions with Approximate Privacy for Applications in Signal Estimation and Learning

Large corporations, government entities and institutions such as hospitals and census bureaus routinely collect our personal and sensitive information for providing services. A key technological challenge is designing algorithms for these services that provide useful results, while simultaneously maintaining the privacy of the individuals whose data are being shared. Differential privacy (DP) is a cryptographically motivated and mathematically rigorous approach for addressing this challenge. Under DP, a randomized algorithm provides privacy guarantees by approximating the desired functionality, leading to a privacy-utility trade-off. Strong (pure DP) privacy guarantees are often costly in terms of utility. Motivated by the need for a more efficient mechanism with better privacy-utility trade-off, we propose Gaussian FM, an improvement to the functional mechanism (FM) that offers higher utility at the expense of a weakened (approximate) DP guarantee. We analytically show that the proposed Gaussian FM algorithm can offer orders of magnitude smaller noise compared to the existing FM algorithms. We further extend our Gaussian FM algorithm to decentralized-data settings by incorporating the CAPE protocol and propose capeFM. Our method can offer the same level of utility as its centralized counterparts for a range of parameter choices. We empirically show that our proposed algorithms outperform existing state-of-the-art approaches on synthetic and real datasets.

50 Years of Pediatric Hemostasis: Knowledge, Diagnosis, and Treatment

Studies from the past 50 years have contributed to the expanding knowledge regarding developmental hemostasis. This is a dynamic process that begins in the fetal phase and is characterized by physiological variations in platelet counts and function, and concentrations of most coagulation factors and the native coagulation inhibitors in early life, as compared with adulthood. The developmental hemostasis studies since the 1980 to 1990s established the laboratory reference values for coagulation factors. It was only a decade or two later, that thromboelastography (TEG) or (rotational thromboelastometry [ROTEM]) as well as thrombin generation studies, provided special pediatric reference values along with the ability to evaluate clot formation and lysis. In addition, global whole blood-based clotting assays provided point of care guidance for proper transfusion support to children hospitalized in intensive care units or undergoing surgery. Although uncommon, thrombosis in children and neonates is gaining increasing recognition, typically as a secondary complication in sick children. Bleeding in children, and particularly intracerebral hemorrhage in newborns, still represent a therapeutic challenge. Notably, our review will outline the advancements in understanding developmental hemostasis and its manifestations, with respect to the pathophysiology of thrombosis and bleeding complications in young children. The changes of transfusion policy and approach to thrombophilia testing during the last decade will be mentioned. Subsequently, a brief summary of the data on anticoagulant treatments in pediatric patients will be presented. Finally, we will point out the 10 most cited articles in the field of pediatric and neonatal hemostasis.

Developing Chinese race-specific warfarin dose prediction algorithms

BACKGROUND

Numerous genotype-guided warfarin dosing algorithms have been developed to individualize warfarin doses, but they can only explain 47-52% of the variability.

AIM

This study aimed to develop new warfarin algorithms suitable to predict the stable warfarin dose for the Chinese population and to compare their prediction performance with those of the most commonly used algorithms.

METHOD

Multiple linear regression analysis with the warfarin optimal dose (WOD), logarithm (log) WOD, 1/WOD, and [Formula: see text], respectively, as the dependent variables were performed to deduce a new warfarin algorithm (NEW-Warfarin). WOD was the stable dose that maintained the international normalized ratio (INR) within the target range (2.0-3.0). Three major genotype-guided warfarin dosing algorithms were selected and compared against NEW-Warfarin predictive performance using the mean absolute error (MAE). Furthermore, patients were divided into five groups according to warfarin indications [atrial fibrillation (AF), pulmonary embolism (PE), cardiac-related disease (CRD), deep vein thrombosis (DVT), and other diseases (OD)]. Multiple linear regression analyses were also performed for each group.

RESULTS

The regression equation with [Formula: see text] as the dependent variable had the highest coefficient of determination (R² = 0.489). The NEW-Warfarin had the best predictive accuracy compared to the three algorithms selected. Group analysis, according to indications, showed that the R² of the five groups were PE (0.902) > DVT (0.608) > CRD (0.569) > OD (0.436) > AF (0.424).

CONCLUSION

Dosing algorithms based on warfarin indications are more suitable for predicting warfarin doses. Our research provides a novel strategy to develop indication-specific warfarin dosing algorithms to improve the efficacy and safety of warfarin prescribing.

Deriving mechanism-based pharmacodynamic models by reducing quantitative systems pharmacology models: An application to warfarin

Quantitative systems pharmacology (QSP) models integrate comprehensive qualitative and quantitative knowledge about pharmacologically relevant processes. We previously proposed a first approach to leverage the knowledge in QSP models to derive simpler, mechanism-based pharmacodynamic (PD) models. Their complexity, however, is typically still too large to be used in the population analysis of clinical data. Here, we extend the approach beyond state reduction to also include the simplification of reaction rates, elimination of reactions, and analytic solutions. We additionally ensure that the reduced model maintains a prespecified approximation quality not only for a reference individual but also for a diverse virtual population. We illustrate the extended approach for the warfarin effect on blood coagulation. Using the model-reduction approach, we derive a novel small-scale warfarin/international normalized ratio model and demonstrate its suitability for biomarker identification. Due to the systematic nature of the approach in comparison with empirical model building, the proposed model-reduction algorithm provides an improved rationale to build PD models also from QSP models in other applications.

Local Ancestry-Informed Candidate Pathway Analysis of Warfarin Stable Dose in Latino Populations

Accuracy of warfarin dose prediction algorithms may be improved by including data from diverse populations in genetic studies of dose variability. Here, we surveyed single nucleotide polymorphisms in vitamin K-related genetic pathways for association with warfarin dose requirements in two admixed Latino populations in standard-principal component adjusted and contemporary-local ancestry adjusted regression models. A total of five variants from vitamin K-related genes/pathways were associated with warfarin dose in both cohorts (P < 0.0125) in standard models. Local ancestry-adjusted analysis unveiled 35 associated variants with absolute effects ranging from β = 9.04 ( ±2.23) to 39.18 ( ±10.89) per ancestral allele in the discovery cohort and β = 6.47 (± 2.02) to 17.82 (± 6.83) in the replication cohort. Importantly, we demonstrate the technical validity of the Tractor model in cohorts with admixed ancestry from three founder populations and bring attention to the technical hurdles obstructing the inclusion of diverse, especially admixed, populations in pharmacogenomic research.

Precision Medicine and the future of Cardiovascular Diseases: A Clinically Oriented Comprehensive Review

Cardiac diseases form the lion's share of the global disease burden, owing to the paradigm shift to non-infectious diseases from infectious ones. The prevalence of CVDs has nearly doubled, increasing from 271 million in 1990 to 523 million in 2019. Additionally, the global trend for the years lived with disability has doubled, increasing from 17.7 million to 34.4 million over the same period. The advent of precision medicine in cardiology has ignited new possibilities for individually personalized, integrative, and patient-centric approaches to disease prevention and treatment, incorporating the standard clinical data with advanced "omics". These data help with the phenotypically adjudicated individualization of treatment. The major objective of this review was to compile the evolving clinically relevant tools of precision medicine that can help with the evidence-based precise individualized management of cardiac diseases with the highest DALY. The field of cardiology is evolving to provide targeted therapy, which is crafted as per the "omics", involving genomics, transcriptomics, epigenomics, proteomics, metabolomics, and microbiomics, for deep phenotyping. Research for individualizing therapy in heart diseases with the highest DALY has helped identify novel genes, biomarkers, proteins, and technologies to aid early diagnosis and treatment. Precision medicine has helped in targeted management, allowing early diagnosis, timely precise intervention, and exposure to minimal side effects. Despite these great impacts, overcoming the barriers to implementing precision medicine requires addressing the economic, cultural, technical, and socio-political issues. Precision medicine is proposed to be the future of cardiovascular medicine and holds the potential for a more efficient and personalized approach to the management of cardiovascular diseases, contrary to the standardized blanket approach.

Development and validation of a bleeding risk prediction score for patients with mitral valve stenosis and atrial fibrillation or mechanical heart valves receiving long-term warfarin therapy

AIMS

This study aimed to develop and validate a new bleeding risk score to predict warfarin-associated major bleeding for patients with mitral valve stenosis with atrial fibrillation (MSAF) or mechanical heart valves (MHV).

METHODS

A multicentre, retrospective cohort study was conducted at 3 hospitals in Thailand. Adult patients with MSAF or MHV receiving warfarin for ≥3 months during 2011-2015 were identified. Data collection and case validation were performed electronically and manually. Potential variables were screened using the least absolute shrinkage and selection operator. Multivariate logistic regression analysis using stepwise backward selection was used to construct a risk score. Predictive discrimination of the score was evaluated using the C-statistic. Calibration was assessed using the Hosmer-Lemeshow goodness-of-fit test.

RESULTS

There were 1287 patients (3903.41 patient-year of follow-up), with 192 experiencing bleeding (4.92 event/100 patient-year) in the derivation cohort. A new bleeding risk score termed, the HEARTS-60 + 3 score (hypertension/history of bleeding; external factors, e.g., alcohol/drugs [aspirin or nonsteroidal anti-inflammatory drugs]; anaemia/hypoalbuminaemia; renal/hepatic insufficiency; time in therapeutic range of <60%; stroke; age ≥60 y; target international normalized ratio of 3.0 [2.5-3.5]), was developed and showed good predictive performance (C-statistic [95% confidence interval] of 0.88 [0.85-0.91]). In the external validation cohort of 832 patients (2018.45 patient-year with a bleeding rate of 4.31 event/100 patient-year), the HEARTS-60 + 3 score showed a good predictive performance with a C-statistic (95% confidence interval) of 0.84 (0.81-0.89).

CONCLUSION

The HEARTS-60 + 3 score shows a potential as a bleeding risk prediction score in MSAF or MHV patients.

Evaluation of supervised machine learning algorithms in predicting the poor anticoagulation control and stable weekly doses of warfarin

BACKGROUND

Machine learning algorithms (MLAs) carry a huge potential in identifying predicting factors and are being explored for their utility in the field of personalized medicine.

AIM

We aimed to investigate MLAs for identifying predictors (clinical and genetic) of poor anticoagulation status (ACS) and stable weekly warfarin dose (SWWD).

METHOD

Clinical factors, in addition to the CYP2C9, VKORC1, and CYP4F2 genotypes, were obtained for patients receiving warfarin for at least the previous six months. The C5.0 decision tree classification algorithm was used to predict poor ACS while classification and regression tree analysis (CART), in addition to the Chi-square automatic interaction detector (CHAID), was used to predict SWWD. The percentage of patients within 20% of the actual dose, root mean squared error (RMSE), and area under the receiver-operating characteristics curve (AUROC) were identified as performance indicators of the models.

RESULTS

In the C5.0 classification decision tree, the CYP4F2 genotype was the strongest predictor of ACS (AUROC = 0.53). In the CART analysis of SWWD, VKORC1 polymorphisms were the most significant predictor, followed by the CYP2C9 genotype (percentage of patients within 20% of the actual dose = 38.2%, RMSE = 13.6). For the CHAID algorithm, the percentage of patients within 20% of the actual dose was 49%, while the RMSE was found to be 13.4.

CONCLUSION

Genetic and non-genetic predictive factors were identified by the MLAs for ACS and SWWD. Further, the need to externally validate the MLAs in a prospective study was highlighted.

Automated warfarin dose prediction for Asian, American, and Caucasian populations using a deep neural network

Existing warfarin dose prediction algorithms based on pharmacogenetics and clinical parameters have not been used clinically due to the absence of external validation, lack of assessment for clinical utility, and high risk of bias. Moreover, given the high degree of heterogeneity across different datasets used to develop these algorithms, it is unsurprising that prediction errors remain high, and dosing accuracy is dependent on specific ethnic populations. To circumvent these challenges, deep neural models are increasingly used to improve the precision and accuracy of warfarin dose predictions. Hence, this study sought to develop a deep learning-based model using a well-established curated dataset of over 6000 patients from the International Warfarin Pharmacogenomics Consortium (IWPC). Clinically-relevant input data such as physical attributes, medical conditions, concomitant medications, genotype status of functional warfarin genetic polymorphisms, and therapeutic INR were entered followed by applying a unique and robust training and validation method. The deep model yielded a low average mean absolute error (MAE) of 7.6 mg/week and a relatively low mean percentage of error of 40.9% in Asians, 14.2 mg/week MAE and 36.9% in African Americans, and 12.7 mg/week MAE and 45.4% mean percentage of error in White Caucasians. This model also resulted in 36.4% of all patients with a predicted dose within 20% of the administered dose. Hence, our proposed deep model provides an alternative to predicting warfarin dose in the clinical setting upon validation in ethnically-similar datasets.

Control-Relevant Adaptive Personalized Modeling From Limited Clinical Data for Precise Warfarin Management

Warfarin is a challenging drug to administer due to the narrow therapeutic index of the International Normalized Ratio (INR), the inter- and intra-variability of patients, limited clinical data, genetics, and the effects of other medications. Goal: To predict the optimal warfarin dosage in the presence of the aforementioned challenges, we present an adaptive individualized modeling framework based on model (In)validation and semi-blind robust system identification. The model (In)validation technique adapts the identified individualized patient model according to the change in the patient's status to ensure the model's suitability for prediction and controller design. Results: To implement the proposed adaptive modeling framework, the clinical data of warfarin-INR of forty-four patients has been collected at the Robley Rex Veterans Administration Medical Center, Louisville. The proposed algorithm is compared with recursive ARX and ARMAX model identification methods. The results of identified models using one-step-ahead prediction and minimum mean squared analysis (MMSE) show that the proposed framework effectively predicts the warfarin dosage to keep the INR values within the desired range and adapt the individualized patient model to exhibit the true status of the patient throughout treatment. Conclusion: This paper proposes an adaptive personalized patient modeling framework from limited patientspecific clinical data. It is shown by rigorous simulations that the proposed framework can accurately predict a patient's doseresponse characteristics and it can alert the clinician whenever identified models are no longer suitable for prediction and adapt the model to the current status of the patient to reduce the prediction error.

A case of pulmonary embolism with bad warfarin anticoagulant effects caused by E. coli infection

Warfarin is an anticoagulant commonly used as an oral drug in preventing and treating thromboembolic diseases. The international normalized ratio (INR) is a clinical monitoring anticoagulation intensity index that adjusts the dose based on important references. In particular, INR value must be strictly monitored when warfarin is used for anticoagulation therapy in infected patients. Herein, we report a 54-year-old female patient diagnosed with pulmonary embolism and venous thrombosis of the lower limbs. After the warfarin administration, the INR was always substandard. The patient did not take other warfarin-interacting drugs or foods during the hospital stay. Metagenome next-generation sequencing suggested a bloodstream infection caused by Escherichia coli, which was further confirmed by blood culture. After meropenem administration for anti-infective treatment, the INR value rose rapidly to a standard level. Considering the lack of relevant reports, this case is the first report of potential interaction between E. coli and warfarin. Further, in patients with thromboembolic diseases complicated by infection, antibiotics should be chosen reasonably with close monitoring of the INR to avoid the interaction of warfarin and antibiotics and to ensure the effectiveness and safety of warfarin treatment.

ABO blood group antigens and differential glycan expression: Perspective on the evolution of common human enzyme deficiencies

Enzymes catalyze biochemical reactions and play critical roles in human health and disease. Enzyme variants and deficiencies can lead to variable expression of glycans, which can affect physiology, influence predilection for disease, and/or directly contribute to disease pathogenesis. Although certain well-characterized enzyme deficiencies result in overt disease, some of the most common enzyme deficiencies in humans form the basis of blood groups. These carbohydrate blood groups impact fundamental areas of clinical medicine, including the risk of infection and severity of infectious disease, bleeding risk, transfusion medicine, and tissue/organ transplantation. In this review, we examine the enzymes responsible for carbohydrate-based blood group antigen biosynthesis and their expression within the human population. We also consider the evolutionary selective pressures, e.g. malaria, that may account for the variation in carbohydrate structures and the implications of this biology for human disease.

Pharmacogenomic variabilities in geo-ancestral subpopulations and their clinical implications: Results of collaborations with Hmong in the United States

Underrepresentation of subpopulations within geo-ancestral groups engaged in research can exacerbate health disparities and impair progress toward personalized medicine. This is particularly important when implementing pharmacogenomics which uses genomic-based sources of variability to guide medication selection and dosing. This mini-review focuses on pharmacogenomic findings with Hmong in the United States and their potential clinical implications. By actively engaging Hmong community in pharmacogenomic-based research, several clinically relevant differences in allele frequencies were observed within key pharmacogenes such as CYP2C9 and CYP2C19 in Hmong compared to those in either East Asians or Europeans. Additionally, using state-of-the-art genome sequencing approaches, Hmong appear to possess novel genetic variants within CYP2D6, a critical pharmacogene affecting pharmacokinetics of a broad range of medications. The allele frequency differences and novel alleles in Hmong have translational impact and real-world clinical consequences. For example, Hmong patients exhibited a lower warfarin stable dose requirement compared to East Asian patients. This was predicted based on Hmong's unique genetic and non-genetic factors and confirmed using real-world data from clinical practice settings. By presenting evidence of the genetic uniqueness and its translational impact within subpopulations, such as the Hmong, we hope to inspire greater inclusion of other geo-ancestrally underrepresented subpopulations in pharmacogenomic-based research.

Optimizing warfarin dosing using deep reinforcement learning

Warfarin is a widely used anticoagulant, and has a narrow therapeutic range. Dosing of warfarin should be individualized, since slight overdosing or underdosing can have catastrophic or even fatal consequences. Despite much research on warfarin dosing, current dosing protocols do not live up to expectations, especially for patients sensitive to warfarin. We propose a deep reinforcement learning-based dosing model for warfarin. To overcome the issue of relatively small sample sizes in dosing trials, we use a Pharmacokinetic/ Pharmacodynamic (PK/PD) model of warfarin to simulate dose-responses of virtual patients. Applying the proposed algorithm on virtual test patients shows that this model outperforms a set of clinically accepted dosing protocols by a wide margin. We tested the robustness of our dosing protocol on a second PK/PD model and showed that its performance is comparable to the set of baseline protocols.

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.