Abstract Warfarin, an oral anticoagulant, is widely used to prevent and treat thromboembolic events, including non-valvular atrial fibrillation (NVAF). According to its narrow therapeutic index, the international normalized ratio (INR) is used to monitor efficacy and safety. The INR outside the target range leads to bleeding or thromboembolic complications, establishing dose optimization of warfarin therapy as challenging. Determining the appropriate warfarin dose requires a comprehensive understanding of the factors influencing dose variability. This study aims to identify the factors affecting the INR and warfarin dosing. A cross-sectional study was conducted among all patients receiving warfarin therapy for NVAF, who were followed at Phanthong Hospital, Thailand. Descriptive statistics with regression were used to investigate the influence on the INR and warfarin dosing. All 104 patients, 61 (58.7%) females with a mean age of 73.6 years, were included in the study. The medians of weekly warfarin doses and the INR were 19.5 mg and 2.14, respectively. The weekly warfarin dose was the factor affecting the INR. For multiple analysis, only estimated glomerular filtration rate (eGFR) was significantly associated with the warfarin dosing provided to maintain within the therapeutic INR (P-value < 0.05). The predicted weekly warfarin dose could use the equation of: Dose in mg = (0.17 x eGFR) + (6.7 x INR) – 6.6. Although INR-based treatment of warfarin follows the Thai guideline, renal function-based warfarin dosing could help patients maintain within the therapeutic INR. Furthermore, this factor could be applied in a pharmacodynamic study to optimize time in the therapeutic range.

Keywords: Warfarin, International normalized ratio, Non-valvular atrial fibrillation, Optimal dosing, Estimated glomerular filtration rate

Citation:  Chaipatiwat, W., Tiamsati, P., Orakul, K., and Chaivichacharn, P. 2026. Factors affecting international normalized ratio and warfarin individualized dosing in patients with non-valvular atrial fibrillation at Phanthong hospital, Thailand. Natural and Life Sciences Communications. 25(4): e2026073.

Graphical Abstract:

INTRODUCTION

Warfarin is frequently prescribed as an oral anticoagulant for the prevention and treatment of thromboembolic disorders, particularly in patients with non-valvular atrial fibrillation (NVAF) having an increased risk of ischemic stroke (Van Gelder et al., 2024). Nevertheless, the efficacy and safety of warfarin therapy were clinical challenges due to its narrow therapeutic index and considerable interpatient variability in dosage requirements (Cho et al., 2007; Joglar et al., 2024; Van Gelder et al., 2024). Monitoring of warfarin use is done using the international normalized ratio (INR) (Van Gelder et al., 2024). Subtherapeutic and supratherapeutic INR led to increased risks of thromboembolism and bleeding, respectively (Priksri et al., 2019; Van Gelder et al., 2024). Interpatient variability in dosage requirements complicated therapy and could be explained by several factors, including age, body weight, dietary vitamin K intake, comedications, comorbidities, and genetic factors (Tham et al., 2006; Miao et al., 2007; Gage et al., 2008; The International Warfarin Pharmacogenetics Consortium, 2009; Lurie et al., 2010; Chumnumwat et al., 2018; Pongbangli et al., 2019).

Recently, renal function has been recognized as a crucial factor influencing warfarin dosing (Yaengkratok and Boonmuang, 2020; Ning et al., 2021). Patients with renal impairment might be suggested to reduce the dosing (Limdi et al., 2010; Ichihara et al., 2015), while data on the relationship between renal function and warfarin dose requirements in Thai populations remain limited (Yaengkratok et al., 2022), particularly within secondary care hospitals. Clinical practice guidelines in Thailand provided standard protocols, which still had only the INR-based dosing, for warfarin initiation and monitoring (The Heart Association of Thailand, 2010). Therefore, this study aims to investigate the factors affecting the INR and to find the optimal weekly warfarin dose in NVAF patients treated at a secondary hospital in Thailand, which could be applicable in real-world clinical settings.

MATERIALS AND METHODS

Study design and setting

This was a cross-sectional study conducted at Phanthong Hospital, where is a public provincial hospital in Chonburi, Thailand. Data were collected in the outpatient-based warfarin clinic from March to May 2025. The study protocol was approved by the Institutional Review Board and the Human Research Ethics Committee of Chonburi Provincial Public Health Office (Protocol No. 017/2568).

Study population

The eligible patients who were 18 years or over, had a confirmed diagnosis of NVAF, understood the Thai language, and had continuous warfarin treatment at steady state were included in this study. However, patients who were pregnant, breastfeeding, had self-discontinued warfarin for more than 7 days, or had a history of discontinuing warfarin for less than 7 days at least 3 consecutive times were excluded.

Data collection

A clinical pharmacist collected data on the routine workflow of the warfarin clinic, which has pre- and post-counseling by pharmacists. Demographics and clinical data consisting of age, body weight, height, gender, ethnicity, smoking status, alcohol use, serum creatinine (SCr), calculated creatinine clearance (CrCl) using the Cockcroft-Gault equation, estimated glomerular filtration rate (eGFR) using the CKD-EPI equation, hemoglobin A1C, CHA₂DS₂-VASc score, hemoglobin, hematocrit, platelet, comedications, and comorbidities (e.g., hypertension, dyslipidemia, diabetes mellitus type II, chronic heart failure, ischemic heart disease, stroke, hypothyroidism, hyperthyroidism, and gout) were extracted for each patient. Furthermore, the weekly dosing and duration of receiving warfarin with the monitored INR were collected.

Statistical analysis

All data were entered into the analysis, and the missing data were discretized, but if it had more than 10%, model-based imputation was used. Descriptive statistics, including frequencies with percentages, mean ± standard deviation, or median (interquartile range: IQR), were used to summarize patient characteristics. In line with hierarchical testing, relationship screenings of characteristics were tested by Pearson correlation. Sample size adequacy for the multivariable analysis was considered based on at least 10 observations per candidate predictor (Hair et al., 2019). Linear regression was used to find the factors affecting INR, using STATA software version 18.5 BE.

The INR between 2.0 and 3.0 was defined as the target range. Consequently, data from patients having INR within the target range were used to determine the factors associated with the warfarin weekly dosing. The significant factors from the multiple analysis providing the INR within the target range were used as input into the predicting model with INR, which was modeled.

For model evaluation, the appropriateness of the predicting model was assessed by graphical analysis, and the reliability was determined by the bootstrapping approach, which performed 1,000 replicates under sampling with replacement, providing a 95% confidence interval (95%CI). Following confirmation of adequate performance, the predictive equation was considered appropriate for clinical application within the studied population.

RESULTS

Characteristics

This study included 104 Thai patients, 61 (58.7%) females, in the analysis, and their characteristics are demonstrated in Table 1. The elderly, with a normal-distributed mean age of 73.6 ± 10.5 years, were the most common patients receiving warfarin. Patients presented with multiple comorbidities, with the most being hypertension (85.6%), followed by dyslipidemia (43.3%) and diabetes mellitus (40.4%). These patients had the trend of high risk of stroke, indicated by the high CHA₂DS₂-VASc score.

According to the 45 (43.3%) and 15 (14.4%) patients who had subtherapeutic and supratherapeutic INR, respectively, there was a direction of INR levels below target over INR levels above target in the context of real-world secondary care. For 13 patients having a previous stroke, 6 (46.2%) patients carrying subtherapeutic INR had a higher occurrence than therapeutic INR (30.8%) and supratherapeutic INR (23.1%).

Table 1. Patient characteristics.

Note: N: number; Duration: duration of warfarin therapy; Dose: warfarin dose; COPD: chronic obstructive pulmonary disease; INR: international normalized ratio; SCr: serum creatinine; BMI: body mass index; CrCl: creatinine clearance; eGFR: estimated glomerular filtration rate; HbA1C: hemoglobin A1C; IQR: interquartile range.

Factors affecting INR

For covariate screening, only the weekly warfarin dose was significantly correlated with INR. Thus, the regression analysis of factors affecting INR found that only a weekly warfarin dose was significantly related to INR (coefficient = -0.017, constant = 2.6, P-value = 0.047). It is widely accepted that the weekly warfarin dose should be adjusted by INR. According to the Thai guideline protocol, the weekly dosing, with its wide range of changes, was adjusted by the current INR of patients, as demonstrated in Table 2 (The Heart Association of Thailand, 2010). Therefore, the alteration of the weekly dosing among patients achieving the target INR with factors should be detected.

Table 2. Warfarin dosing protocols to attain a target international normalized ratio by the Thai guidelines for the treatment of patients with oral anticoagulants.

Note: INR: international normalized ratio; Vit.K1: vitamin K1; IV: intravenous; q 12 h: every 12 hours; FFP: fresh frozen plasma.

Factors affecting warfarin weekly dose among patients whose INR were within the range

For patients having INR within the target range in this study, the screening by Pearson correlation found 3 significant covariates consisting of age, eGFR, and CrCl (P-value = 0.022, 0.012, and 0.020, respectively), which were used to test as factors influencing the weekly warfarin dose. The relationship between weekly warfarin dose and each factor did not show a non-linear function, as demonstrated in Figure 1. The INR did not show a significant relationship with the weekly warfarin dose (P-value = 0.568). For multiple regression analysis, including INR into the eGFR-based model did not improve model fit (P-value = 0.192). Only the eGFR was a significant factor, influencing the weekly warfarin dose required to maintain INR in the target range. The results are demonstrated in Table 3. Although the routine treatment with a full multidisciplinary team following the Thai guidelines for the treatment of patients with oral anticoagulants could manage and omit the predictive factors affecting INR, there remained a chance that patients would fall outside the target INR range.

Figure 1. Scatter plots demonstrating the relationship of weekly warfarin dose (Y-axis) and significant factors (X-axis).

Table 3. Regression analysis for factors affecting weekly warfarin dose, providing patients maintain international normalized ratio in the target range.

Note: INR: international normalized ratio; CrCl: calculated creatinine clearance; eGFR: estimated glomerular filtration rate; β: beta coefficient; SE: standard error; R²: coefficient of determination. *Age at the time of visiting the warfarin clinic.

Models for warfarin dosing and model evaluation

The optimal dosing was modeled by integrating INR with a significant factor of the eGFR, and a predicting model for weekly warfarin dose was:

Dose in mg = (0.17 x eGFR) + (6.7 x target INR) – 6.6

For clinical interpretability, this weekly-dose equation could be translated into a practical daily regimen, targeting the INR of 2.5. Consistent with current guideline recommendations, an initial dose of 3 mg/day is proposed for general patients (The Heart Association of Thailand, 2010). Based on the derived model, dose reduction is recommended for patients with impaired renal function, with a daily dose of 2.5 mg/day for those with stage 3 renal impairment (eGFR 30–59 mL/min/1.73 m²) and 2.0 mg/day for those with stage 4 renal impairment (eGFR 15–29 mL/min/1.73 m²).

The predicting model was plotted as a fitted plot and a residual plot, which are shown in Figure 2. The fitted plot showed the linear function between observed and predicted values of warfarin weekly dose. The fitted line of the predicting model was consistently overlaid with the scattering plot of observed warfarin weekly dose, and the model fit of the predicting model was accepted. Parallelly, the residual plot obtained from the residuals and fitted values of the warfarin weekly dose emerged as a homogeneous distribution around the zero line of residuals. Thus, the model misspecification was not detected or shown in the systematic trend.

The bootstrapping approach provided the estimates, which were comparable to the values of the predicting model, as exhibited in Table 4. These results confirmed the reliability of the predicting model. Collectively, these findings indicated that the model demonstrated adequate performance within the studied population.

Figure 2. The goodness-of-fit plots of the predicting model. The plot with its fitted dashed line shows the observed versus predicted warfarin weekly dose (A), and the plot with its fitted dashed line shows the residuals versus predicted warfarin weekly dose (B).

Table 4. Bootstrapping analysis (1,000 replicates).

Note: eGFR: estimated glomerular filtration rate; INR: international normalized ratio; β: beta coefficient; SE: standard error; 95%CI: 95% confidence interval (The 95%CI consisted of the 2.5^th and 97.5^th percentiles.).

DISCUSSION

This study investigated the factors affecting INR and the optimal weekly warfarin dosing among Thai patients with NVAF. The main findings indicated that while no single factor considerably forecasted INR change in real-world practice, the eGFR emerged as a significant predictor of the required weekly warfarin dose for patients achieving the therapeutic INR. The absence of predictors for INR change could reflect the standard of care at the warfarin clinic in the secondary care hospital, where a multidisciplinary team actively manages patients following national guidelines. This intensive monitoring and patient counseling likely diminished the influence of several individual factors (Tajai et al., 2018). Nonetheless, the observation that over half of the patients (57.7%) still had INR values outside the target range highlighted the inherent trouble in warfarin management and emphasized the need for more precision dosing tools.

The distribution of INR values in this study required additional determination. The observation that a higher proportion of patients were subtherapeutic INR (43.3%) rather than supratherapeutic INR (14.4%) provided important insights into the practical challenges encountered at the secondary care hospital. This trend indicated a conservative dosing strategy by clinicians, who could prioritize the avoidance of immediate and noticeable bleeding events over the more delayed risk of thromboembolism. In a resource-limited setting, resolving major bleeding may become considerably more challenging than managing a subtherapeutic INR. This pattern implied that the patients, on average, were inadequately protected against their elevated inherent stroke risk, as evidenced by the median CHA₂DS₂-VASc score of 4. The finding highlighted an important topic for enhancement in clinical practice, wherein instruments such as the suggested dosing model could enable clinicians to administer dosing with more confidence and efficacy, potentially changing this distribution towards improved therapeutic coverage without elevating the risk of bleeding.

This study introduced a novel investigation regarding renal function-based dosing, which was a topic that remains broadly unexamined in a Thai population. Our investigation indicated that only eGFR, calculated using the CKD-EPI equation, remained a significant predictor of the weekly warfarin dose, while age and CrCl failed. This pattern was compatible with the strong intercorrelation among these variables, suggesting that eGFR captured the relevant effect of age-related and creatinine-based changes in renal function. As renal function declines with increasing age (Kerdchantuk et al., 2010), eGFR was the most informative of the three parameters for dose individualization. In our regression model, which excluded INR due to its lack of significance, the coefficient for eGFR was 0.15, which implied that for every 10 mL/min/1.73 m² decrease in a patient's eGFR, the weekly warfarin dose should be proactively reduced by a clinically meaningful amount of 1.5 mg. This result was consistent with several studies, which could be explained by the liver cytochrome P450 having been down-regulated among patients with renal impairment, resulting in decreased nonrenal clearance (Dreisbach et al., 2003; Déri et al., 2020). In accordance with this, a sparse unchanged form of warfarin was excreted via the kidney, which has the chance for the renal clearance to reflect the reduced dose when abnormal renal function is present (Owen, 2004). Not only pharmacokinetics but also pharmacodynamics, the previous study by Jun et al. (2015) showed that kidney impairment was associated with increased sensitivity to warfarin and a risk of bleeding in elders with atrial fibrillation. 

The impacts of genetic factors clearly described the interindividual variabilities of warfarin dosing with its pharmacokinetics and pharmacodynamics and could explain the individuals' differences in ethnicity (Tham et al., 2006; Miao et al., 2007; Gage et al., 2008; The International Warfarin Pharmacogenetics Consortium, 2009; Sangviroon et al., 2010; Pongbangli et al., 2019). Although the pharmacogenetic-based dosing was recommended in different populations, including Asians such as Japanese (Takahashi et al., 2006), Chinese (Shi and Deng, 2024), and Thais (Sangviroon et al., 2010; Pongbangli et al., 2019; Ujjin et al., 2024), the genetically adjusted dosing still had the factor of renal function in the recommended dosage regimen (Limdi et al., 2010). There showed the confidence of the renal function factor when compensated with genetic factors.

For the Thai population, Ujjin et al. (2024), investigated that the simplified warfarin dosing formula, which was comparable with the fixed dose of 3 mg, had the factors including age, body weight, and chronic heart failure. However, this study did not mention renal function, making the application of simplified dosing for the elderly and patients with renal impairment questionable. Their setting was a referral hospital, which differs from ours. In parallel, Sangviroon et al. (2010) demonstrated that the age-based weekly warfarin dosing integrating with VKORC1 and CYP2C9 was described in Thai patients with the valvular replacements, which was a different indication of NVAF from our study. It was performed in a hospital affiliated with a medical school, but the genetic testing in several secondary care hospitals was not available and had the augmented cost for Thai patients. And the renal function affecting dosing might be masked due to those patients having the average age of 49 years.

Our study investigating Thai patients with NVAF was the first validated model for dosing. The model evaluation affirmed the structural integrity and predictive reliability of the equation. The residual analysis revealed no systematic pattern, thereby meeting the assumption of homoscedasticity. This indicated that the predictive error was constant across all levels of warfarin dosing, avoiding bias at the extremes of the dose range. The bootstrapping analysis additionally confirmed the internal validity. The reliability of the bootstrapped estimates and their confidence intervals to the original model's values demonstrated the stability and reproducibility of the eGFR coefficient. This validation process confirmed the model's robustness and unbiased predictor suitable for clinical application.

This finding is relevant to the ongoing controversy in Thailand regarding the preferential use of eGFR or CrCl for dose adjustments. Our results encouraged the KDIGO 2024 guideline, recommending eGFR over CrCl for estimating renal function (Awdishu et al., 2025) and dose adjustment of warfarin. The CKD-EPI equation is considered more accurate than the Cockcroft-Gault equation for estimating glomerular filtration rate, especially in the elderly, who were predominant in our study. In these elderly patients, renal function decline is physiological (Xu et al., 2024), and estimations from the Cockcroft-Gault equation could be less accurate due to reduced muscle mass (Weinstein and Anderson, 2010). This finding about eGFR from the CKD-EPI equation was confirmed and preferred for guiding therapy in this group. Especially, several recommendations suggested that patients with renal impairment or having eGFR lower than 60 mL/min/1.73 m² could reduce the warfarin dose, which would be reduced by approximately 10% to 20% (Limdi et al., 2009; Limdi et al., 2010; Sakaan et al., 2014). However, the initial dose had never been investigated. This study provided the initial dose based on INR and eGFR.

In addition, the derived dosing model offered a simple and highly practical tool in clinical practice. A major advantage was that the eGFR value was automatically calculated and reported alongside SCr by several hospital laboratories in Thailand. In contrast with CrCl, it required manual calculation using age, body weight, SCr, and gender (Cockcroft and Gault, 1976). These complicated it and led to higher errors of measurement. The direct availability of eGFR allows for the immediate and seamless application of this dosing formula in a routine clinical workflow, potentially helping clinicians select a more appropriate starting or adjusted dose to achieve a therapeutic INR more efficiently.

Furthermore, the simplicity of this eGFR-based model is a major strength for implementation in real-world clinical settings, such as a secondary hospital. Whereas complex pharmacogenetic-based dosing, including CYP2C9 and VKORC1 polymorphisms (Tham et al., 2006; Miao et al., 2007; Sangviroon et al., 2010), could offer higher precision, their use is often limited by high costs and lack of routine availability in Thai healthcare facilities. Although CYP2C9 gene, encoding the CYP2C9 enzyme that primarily metabolizes warfarin (Tuba et al., 2021), could be used to inform dosing, this polymorphism might be low frequency in Asians (Kanjanasilp et al., 2005). Our quantitative guidance was an alternative to the current standard protocol, which relies solely on adjusting doses after an INR is found to be out of range. It allowed for personalized dosing and ease of use, making individualized warfarin therapy more accessible.

The presence of 3 outliers in Figure 1, which were intentionally included in the analysis to reflect real-world clinical heterogeneity, warrants specific discussion. The first patient had a comorbidity of pulmonary embolism, inducing hypercoagulability that possibly required a higher warfarin dose to attain the therapeutic INR. The second could potentially be explained by dietary interaction, and the last could not be associated with observable clinical or food/supplement factors. These individual cases highlighted a crucial point. While the eGFR model provided an invaluable and practical baseline for dosing, it could not account for all sources of variability. They highlighted that true therapeutic optimization remains a multifactorial challenge, reinforcing the indispensable role of the clinician in integrating model-based guidance with a comprehensive assessment of the individual patient factors.

The important consideration for clinical application was that this model was not developed in patients with hepatic impairment. Patients with liver disease presented a unique challenge, as their baseline INR might be elevated due to impaired synthesis of vitamin K-dependent clotting factors, independent of warfarin's effect. Furthermore, since warfarin is primarily metabolized by the liver, hepatic dysfunction could dramatically reduce its clearance, leading to a highly unpredictable anticoagulant response (Qamar et al., 2018). Applying this eGFR-based model to such patients would be inappropriate and potentially dangerous, as it did not account for these complex pharmacokinetic/pharmacodynamic alterations. Therefore, the dosing equation derived from this study should be applied with extreme caution, or not at all, in patients with moderate-to-severe liver disease, whose anticoagulation management requires specialized clinical judgment.

It was imperative to define the boundaries of this model's utility, particularly concerning patients with severe renal impairment or end-stage renal disease (ESRD). The patient population used to develop this model predominantly had mild-to-moderate kidney disease, and extrapolating the linear eGFR-dose relationship to patients with ESRD (e.g., eGFR < 15 mL/min/1.73 m² or on dialysis) was not appropriate. This, combined with an inherently high baseline bleeding risk due to platelet dysfunction in ESRD patients, created a highly unpredictable anticoagulant response. Therefore, this model should not be used in the ESRD patients, where warfarin dosing must be guided by highly cautious, individualized clinical assessment with a primary focus on mitigating bleeding risk.

Although this study did not measure the time in therapeutic range (TTR), which was the marker responding to clinical outcome (Van Gelder et al., 2024), our results led to the hypothesis that the eGFR-guided dosing strategy could lead to a higher TTR and subsequently better clinical outcomes for further studies by enabling clinicians to select a more appropriate dose earlier in the treatment course. This model might have the potential to reduce the time spent on dose titration and minimize INR outside the target range.

Several limitations were noted in this study. The cross-sectional and single-center study may limit the applicability of our findings. Moreover, genetic factors that significantly influence warfarin requirements, including known interethnic variability in polymorphisms such as CYP2C9 and VKORC1, as well as anthropometric differences across ethnic groups, were not assessed in this study. Therefore, the predictive equation developed from this Thai population cannot be extrapolated to other races without further validation. Although the patient adherence in this study had over 90% form inquiry, the residual pill counts could not be performed in this setting due to the limited human resources in Thailand. Future multicenter, prospective studies are warranted to validate this dosing model and to incorporate genetic data to further fine-tune dose prediction.

CONCLUSION

This study established that the eGFR is a key factor for determining the weekly warfarin dose, even though no single factor could predict the INR. Rigorous evaluation confirmed the model's reliability and appropriateness. Incorporating eGFR into clinical practice offers a simple, evidence-based strategy to guide and personalize warfarin therapy.

ACKNOWLEDGEMENT

All authors would like to express their sincere gratitude to the directors and all staff of Phanthong Hospital for their invaluable assistance and support during the data collection period. We are profoundly grateful to all the patients who participated in this study.

AUTHOR CONTRIBUTIONS

Watsit Chaipatiwat: Conceptualization (Lead), Methodology (Lead), Software (Equal), Validation (Lead), Formal Analysis (Lead), Investigation (Equal), Resources (Supporting), Data Curation (Lead), Writing - Original Draft (Lead), Visualization (Equal), Supervision (Equal), Project Administration (Equal); Pataranan Tiamsati: Conceptualization (Lead), Methodology (Lead), Software (Equal), Formal Analysis (Lead), Investigation (Supporting), Resources (Supporting), Data Curation (Supporting), Writing - Review & Editing (Equal), Visualization (Lead), Supervision (Supporting), Project Administration (Lead); Kantaphol Orakul: Conceptualization (Equal), Methodology (Equal), Investigation (Lead), Resources (Lead), Data Curation (Supporting), Visualization (Supporting), Project Administration (Supporting); Piyawat Chaivichacharn: Conceptualization (Lead), Methodology (Lead), Software (Lead), Validation (Lead), Formal Analysis (Lead), Investigation (Supporting), Resources (Supporting), Data Curation (Lead), Writing - Review & Editing (Lead), Visualization (Lead), Supervision (Lead), Project Administration (Lead).

CONFLICT OF INTEREST

The authors declare that they have no conflicts of interest.

REFERENCES

Awdishu, L., Maxson, R., Gratt, C., Rubenzik, T., and Battistella, M. 2025. KDIGO 2024 clinical practice guideline on evaluation and management of chronic kidney disease: A primer on what pharmacists need to know. American Journal of Health-System Pharmacy. 82(12): 660-671. https://doi.org/10.1093/ajhp/zxaf044

Cho, H.J., Sohn, K.H., Park, H.M., Lee, K.H., Choi, B., Kim, S., Kim, J.S., On, Y.K., Chun, M.R., and Kim, H.J. 2007. Factors affecting the interindividual variability of warfarin dose requirement in adult Korean patients. Pharmacogenomics. 8(4): 329-337. https://doi.org/10.2217/14622416.8.4.329

Chumnumwat, S., Yi, K., Lucksiri, A., Nosoongnoen, W., Chindavijak, B., Chulavatnatol, S., Sarapakdi, A., and Nathisuwan, S. 2018. Comparative performance of pharmacogenetics‐based warfarin dosing algorithms derived from Caucasian, Asian, and mixed races in Thai population. Cardiovascular Therapeutics. 36(2): e12315. https://doi.org/10.1111/1755-5922.12315

Cockcroft, D.W. and Gault, H. 1976. Prediction of creatinine clearance from serum creatinine. Nephron. 16(1): 31-41. https://doi.org/10.1159/000180580

Déri, M.T., Kiss, Á.F., Tóth, K., Paulik, J., Sárváry, E., Kóbori, L., and Monostory, K. 2020. End-stage renal disease reduces the expression of drug-metabolizing cytochrome P450s. Pharmacological Reports. 72(6): 1695-1705. https://doi.org/10.1007/s43440-020-00127-w

Dreisbach, A.W., Japa, S., Gebrekal, A.B., Mowry, S.E., Lertora, J.J., Kamath, B.L., and Rettie, A.E. 2003. Cytochrome P4502C9 activity in end‐stage renal disease. Clinical Pharmacology and Therapeutics. 73(5): 475-477. https://doi.org/10.1016/s0009-9236(03)00015-8

Gage, B.F., Eby, C., Johnson, J., Deych, E., Rieder, M., Ridker, P., Milligan, P., Grice, G., Lenzini, P., and Rettie, A. 2008. Use of pharmacogenetic and clinical factors to predict the therapeutic dose of warfarin. Clinical Pharmacology and Therapeutics. 84(3): 326-331. https://doi.org/10.1038/clpt.2008.10

Hair, J.F. Jr., Black, W.C., Babin, B.J., and Anderson, R.E. 2019. Multivariate data analysis. 8^th ed. Cengage Learning EMEA, Andover, Hampshire, United Kingdom.

Ichihara, N., Ishigami, T., and Umemura, S. 2015. Effect of impaired renal function on the maintenance dose of warfarin in Japanese patients. Journal of Cardiology. 65(3): 178-184. https://doi.org/10.1016/j.jjcc.2014.08.008

Joglar, J.A., Chung, M.K., Armbruster, A.L., Benjamin, E.J., Chyou, J.Y., Cronin, E.M., Deswal, A., Eckhardt, L.L., Goldberger, Z.D., and Gopinathannair, R. 2024. 2023 ACC/AHA/ACCP/HRS guideline for the diagnosis and management of atrial fibrillation: A report of the American College of Cardiology/American Heart Association joint committee on clinical practice guidelines. Journal of the American College of Cardiology. 83(1): 109-279. https://doi.org/10.1016/j.jacc.2023.08.017

Jun, M., James, M.T., Manns, B.J., Quinn, R.R., Ravani, P., Tonelli, M., Perkovic, V., Winkelmayer, W.C., Ma, Z., and Hemmelgarn, B.R. 2015. The association between kidney function and major bleeding in older adults with atrial fibrillation starting warfarin treatment: Population based observational study. BMJ. 350: h246. https://doi.org/10.1136/bmj.h246

Kanjanasilp, J., Preechagoon, Y., Kaewvichit, S., and Richards, R.M.E. 2005. Population pharmacokinetics of phenytoin in Thai epileptic patients. Chiang Mai University Journal of Natural Sciences. 4: 287-297. 

Kerdchantuk, C., Chaiyakum, A., Kessomboon, N., Kanpittaya, J., Vechakama, P., and Sirivong, D. 2010. Self-assessment symptoms and risk factors for chronic kidney disease screening in primary care. Chiang Mai University Journal of Natural Sciences. 9(1): 39-47.

Limdi, N.A., Beasley, T.M., Baird, M.F., Goldstein, J.A., McGwin, G., Arnett, D.K., Acton, R.T., and Allon, M. 2009. Kidney function influences warfarin responsiveness and hemorrhagic complications. Journal of the American Society of Nephrology. 20(4): 912-921. https://doi.org/10.1681/asn.2008070802

Limdi, N.A., Limdi, M.A., Cavallari, L., Anderson, A.M., Crowley, M.R., Baird, M.F., Allon, M., and Beasley, T.M. 2010. Warfarin dosing in patients with impaired kidney function. American Journal of Kidney Diseases. 56(5): 823-831. https://doi.org/10.1053/j.ajkd.2010.05.023

Lurie, Y., Loebstein, R., Kurnik, D., Almog, S., and Halkin, H. 2010. Warfarin and vitamin K intake in the era of pharmacogenetics. British Journal of Clinical Pharmacology. 70(2): 164-170. https://doi.org/10.1111/j.1365-2125.2010.03672.x

Miao, L., Yang, J., Huang, C., and Shen, Z. 2007. Contribution of age, body weight, and CYP2C9 and VKORC1 genotype to the anticoagulant response to warfarin: Proposal for a new dosing regimen in Chinese patients. European Journal of Clinical Pharmacology. 63(12): 1135-1141. https://doi.org/10.1007/s00228-007-0381-6

Ning, X., Kuang, Y., Yang, G., Xie, J., Miao, D., Guo, C., and Huang, Z. 2021. Influence of renal insufficiency on anticoagulant effects and safety of warfarin in Chinese patients: Analysis from a randomized controlled trial. Naunyn-schmiedeberg's Archives of Pharmacology. 394(6): 1275-1283. https://doi.org/10.1007/s00210-020-02037-3

Owen, J.A. 2004. Renal function and drug therapy. Encyclopedia of endocrine diseases. L.J. DeGroot. Amsterdam, Elsevier. 17–25.

Pongbangli, N., Phrommintikul, A., and Wongcharoen, W. 2019. Simplified warfarin dosing formula to guide the initiating dose in Thai patients. Journal of the Medical Association of Thailand. 102(9): 957-961.

Priksri, W., Rattanavipanon, W., Saejear, W., Tanyasaensook, K., Phrommintikul, A., Chulavatnatol, S., and Nathisuwan, S. 2019. Incidence, risk factors, and outcomes of warfarin‐associated major bleeding in Thai population. Pharmacoepidemiology and drug safety. 28(7): 942-950. https://doi.org/10.1002/pds.4781

Qamar, A., Vaduganathan, M., Greenberger, N.J., and Giugliano, R.P. 2018. Oral anticoagulation in patients with liver disease. Journal of the American College of Cardiology. 71(19): 2162-2175. https://doi.org/10.1016/j.jacc.2018.03.023

Sakaan, S.A., Hudson, J.Q., Oliphant, C.S., Tolley, E.A., Cummings, C., Alabdan, N.A., and Self, T.H. 2014. Evaluation of warfarin dose requirements in patients with chronic kidney disease and end‐stage renal disease. Pharmacotherapy: The Journal of Human Pharmacology and Drug Therapy. 34(7): 695-702. https://doi.org/10.1002/phar.1445

Sangviroon, A., Panomvana, D., Tassaneeyakul, W., and Namchaisiri, J. 2010. Pharmacokinetic and pharmacodynamic variation associated with VKORC1 and CYP2C9 polymorphisms in Thai patients taking warfarin. Drug Metabolism and Pharmacokinetics. 25(6): 531-538. https://doi.org/10.2133/dmpk.dmpk-10-rg-059

Shi, K. and Deng, J. 2024. 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. Pharmacogenetics and Genomics. 34(9): 275-284. https://doi.org/10.1097/fpc.0000000000000545

Tajai, P., Warinruk, J., and Sunkwan, T. 2018. The effects of pharmacist’s counseling on warfarin-related clinical outcomes in pharmacy’s ambulatory care. Interprofessional Journal of Health Sciences. 16(2): 102-108. 

Takahashi, H., Wilkinson, G.R., Nutescu, E.A., Morita, T., Ritchie, M.D., Scordo, M.G., Pengo, V., Barban, M., Padrini, R., and Ieiri, I. 2006. Different contributions of polymorphisms in VKORC1 and CYP2C9 to intra- and inter-population differences in maintenance dose of warfarin in Japanese, Caucasians and African-Americans. Pharmacogenetics and Genomics. 16(2): 101-110. https://doi.org/10.1097/01.fpc.0000184955.08453.a8

Tham, L.S., Goh, B.C., Nafziger, A., Guo, J.Y., Wang, L.Z., Soong, R., and Lee, S.C. 2006. A warfarin‐dosing model in Asians that uses single‐nucleotide polymorphisms in vitamin K epoxide reductase complex and cytochrome P450 2C9. Clinical Pharmacology and Therapeutics. 80(4): 346-355. https://doi.org/10.1016/j.clpt.2006.06.009

The Heart Association of Thailand. 2010. Clinical practice guideline for the management of patients receiving oral anticoagulants. The Heart Association of Thailand Under the Royal Patronage.

The International Warfarin Pharmacogenetics Consortium. 2009. Estimation of the warfarin dose with clinical and pharmacogenetic data. New England Journal of Medicine. 360(8): 753-764. https://doi.org/10.1056/nejmoa0809329

Tuba, S., Ikawati, Z., and Mustofa. 2021. The frequencies of allele distribution of CYP2C9 and CYP2C19 gene polymorphisms in healthy Papuan population, Indonesia. Chiang Mai University Journal of Natural Sciences. 20(3): e2021066. https://doi.org/10.12982/CMUJNS.2021.066

Ujjin, A., Wongcharoen, W., Suwanagool, A., and Chai-Adisaksopha, C. 2024. Optimal strategies to select warfarin dose for Thai patients with atrial fibrillation. Journal of Clinical Medicine. 13(9): 2675. https://doi.org/10.3390/jcm13092675

Van Gelder, I.C., Rienstra, M., Bunting, K.V., Casado-Arroyo, R., Caso, V., Crijns, H.J., De Potter, T.J., Dwight, J., Guasti, L., and Hanke, T. 2024. 2024 ESC Guidelines for the management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS) Developed by the task force for the management of atrial fibrillation of the European Society of Cardiology (ESC), with the special contribution of the European Heart Rhythm Association (EHRA) of the ESC. Endorsed by the European Stroke Organisation (ESO). European Heart Journal. ehae 176. https://doi.org/10.1093/eurheartj/ehae176

Weinstein, J.R. and Anderson, S. 2010. The aging kidney: Physiological changes. Advances in Chronic Kidney Disease. 17(4): 302-307. https://doi.org/10.1053/j.ackd.2010.05.002

Xu, L., Yu, C., Chen, A., Li, C., and Mao, Y. 2024. Longitudinal analysis of renal function changes in elderly populations: Health status evaluation and risk factor assessment. Clinical Interventions in Aging. 19: 1217-1224. https://doi.org/10.2147/cia.s450388

Yaengkratok, S. and Boonmuang, P. 2020. Effect of chronic kidney disease on warfarin use. Thai Bulletin of Pharmaceutical Sciences. 15(2): 109-119. https://doi.org/10.69598/tbps.15.2.109-119

Yaengkratok, S., Pongchaidecha, M., Santimaleeworagun, W., and Boonmuang, P. 2022. Effect of chronic kidney disease on warfarin responsiveness among Thai patients. The Thai Journal of Pharmaceutical Sciences. 46(3): 341-345. https://doi.org/10.56808/3027-7922.2579

OPEN access freely available online

Natural and Life Sciences Communications

Chiang Mai University, Thailand. https://cmuj.cmu.ac.th