Warfarin, a coumarin derivative oral anticoagulant, is widely prescribed globally for the prevention of venous thromboembolism and thromboembolic stroke in patients with either Non-Valvular Atrial fibrillation (AF) or chronic atrial fibrillation [1] [2]. A study has shown that warfarin initiation period bears the highest risk of poor anticoagulation, although long-term warfarin management also has higher chances of over anticoagulation [3]. That is why it is essential to monitor the warfarin initiation and management doses which may be affected due to many reasons either clinical or non-clinical.

Warfarin dose adjustment is based on regular monitoring of international normalized ratio (INR), which is required to reduce the risk of adverse events. Very high or low doses of warfarin may lead to poor INR control as Okumura et al. (2011) have reported that higher warfarin doses are associated with poorer INR control whereas Palaretti et al. (2005) describe that lower warfarin doses are associated with poorer INR control [4] [5]. Henceforth, monitoring of warfarin efficacy and its complications is essential for the management of patients on warfarin therapy. Accurate INR monitoring is essential for optimal anticoagulation management. It is recommended by the professional society guidelines that INR needs to monitor in 4 weeks’ intervals or once every month [6].

Time in therapeutic range (TTR) has been widely used to measure the quality of INR control as it reflects the percentage of time within a therapeutic INR range [7]. A minimum target of TTR ≥ 65% is recommended by the experts to ensure the effectiveness and safety of warfarin [8]. The general INR range recommended by the Western and recently published Malaysian guidelines is 2.0 - 3.0 for all age groups. In addition, an INR range of 2.5 - 3.5 is recommended for a patient with high thrombotic risk, for example, patients with mechanical prosthetic valves [9] [10]. On the other hand, Japanese Circulation recommended different INR ranges according to the age group. An INR range of 2.0 - 3.0 is used for patients < 70 years and a lower INR range of 1.6 - 2.6 is being used by the Japanese physicians for patients ≥ 70 years old. The J-RHYTHM Registry has shown that Japanese physicians tend to go with lower INR range than the usual 2.0 - 3.0 therapeutic range for 60% of patients aged < 70 years old [11].

There is still debate on the optimal INR range for Asian population due to limited evidence. Moreover, several factors affect the patients from achieving and maintaining therapeutic INR range, and these factors could be clinical and non-clinical. In this review paper, we analyzed how the non-clinical factors like socio-demographic, modifiable risk and protective factors affect the international normalized ratio (INR) control in patients who are receiving treatment with warfarin.

1) Age: There is clear evidence on the benefits of warfarin for stroke and thromboembolism prevention in patients with Atrial fibrillation (AF), and the prevalence of AF in patients > 70 years old and >85 years old is >5% and >10% respectively. Although the prevalence of AF is high, the concern on the risk of bleeding has limited these elderly patients from getting optimal anticoagulation prophylaxis [12] [13]. Cross-sectional study design though limits the establishment of causality but has shown that advanced age, especially elderly patients, have the highest risk of poor INR control which increases warfarin-related complications [14].

Advanced age is a non-modifiable risk factor of bleeding; hence it is essential to monitor the dose of warfarin [15]. Shendre et al. (2018) reported that the incidence rate of haemorrhage is significantly highest among the elderly patients on warfarin therapy compared to the young warfarin consumers [Incidence Rate Ratio (IRR) 1.8, 95% Confidence Intervals (CI): 1.12 - 3.0]. Although these elderly patients can achieve good INR control, they still have an increased risk of significant bleeding, which may happen due to either poor hypertension control or age-related frailty [16]. Cohen et al. (2019) described that 8% of their older study population has supra-therapeutic INR level with consequences of higher bleeding rate, a longer length of stay and elevated mortality rate. Approximately 33% of patients had INR overshoot of >3 after warfarin initiation, and 8% of these patients had INR elevations of ≥5 [17].

*NA = Not Applicable.

Regarding time in therapeutic range (TTR), the Veterans Affairs Study to Improve Anticoagulation (VARIA) reported that elderly patients have better TTR level [18]. Similarly, Shendre et al. (2018) described that elderly patients (>70 years) have significantly higher TTR compared to younger warfarin user (p < 0.0001). However, the risk of major bleeding is still higher among older patients despite achieving better INR control [16], which is because as there is no association between high TTR (>70%) and warfarin adverse events [19].

A study conducted in Australia also found that patients < 60 years have low TTR [20]. On the other hand, Abohelaika et al. (2016) demonstrated that increasing age of >70 years is associated with lower TTR of ≤65% [21]. Also, Numao et al. (2017) reported that older age is associated with increased INR variability, which highlights the increased risk of poor anticoagulation [22]. The TTR value is also low in older age [23] [24]. The possible cause could be the convenience of polypharmacy, low body mass index (BMI) and renal or liver dysfunction that may lead to unstable pharmacokinetics [25].

In contrast, several studies have shown that younger age < 45 years or <60 years is associated with poor INR control [26] [27] [28]. Abohelaika et al. (2016) reported that patients < 60 years have poorer INR control compared to elderly patients, and this could be due to their employment issue, alcohol abuse, poor compliance and comorbidities [21]. In contrast, Prochaska et al. (2019) found that age is not a risk factor of sub-therapeutic control [29].

Age has been integrated into a few risk scores of bleeding. Anyhow, this incorporation has only able to improve the predicting ability to bleed moderately. Different risk scores use different age cut-off, which makes it harder to set the age threshold for determining the risk of bleeding [30]. Worsening of normal body function due to age-related frailty could accelerate the adverse events in the elderly [31]. Maximum effectiveness cannot be achieved with a lower INR target range of 1.6 - 2.6 in elderly patients (>70 years), although it might decrease the bleeding risk [32]. Therefore, current evidence recommended the INR target of 2.0 to 3.0 in the elderly as the benefits still outweigh the risk [33].

2) Gender: Various studies identified the associations between gender and INR control. Marie et al. (2012) found that female gender is highly associated with excessive anticoagulation whereas Martin-Perez et al. (2019) did not find any statistically significant association between gender of the patients and INR control [34] [35]. Prochaska et al. (2019) reported that female is associated with sub-therapeutic INR control which is in line with an epidemiology report released in the United States by Rose et al. (2009) [29] [36]. The possible speculated explanation for this gender-specific variation could be due to non-compliance, poor dietary vitamin K intake, different physiological response to warfarin or treatment variation by clinicians [29] [37]. Concerning time in therapeutic range (TTR), the female has been found to be more vulnerable than male as they act as an independent predictor of poor TTR [26] [38].

Abohelaika et al. (2016) reported that females have lower mean TTR by 1.3% and they spent marginally more time below the therapeutic range compared to males; though there is no significant difference in mean time spends beyond the INR target between different genders. More women are monitored domiciliary than men which could be a contributing factor of this TTR variation [21]. In contrast, Bernaitis et al. (2017), Gallagher et al. (2011) and Okumura et al. (2011) reported that there is no association between gender difference and TTR [4] [20] [39].

3) Ethnicity: The warfarin maintenance doses, which required to attain the optimal INR control significantly, are different among patients from different ethnicities (African American > White > Asian) [40]. A study described that African American patients have a lower TTR and an increased risk of bleeding compared to European American patients; this suggests that INR control could be affected by racial difference [41]. Yong et al. (2016) reported that black patients have lower first-year and long-term TTR compared to all other races in the study despite similar INR monitoring frequency intensity. They found different mean TTR among different racial groups. The white racial group had the highest mean TTR after accounting their socio-demographic, clinical and site-level factors. The study also found that the difference in anticoagulation control among different racial groups is not associated with the accessibility to medical facilities as the blacks stay near to the medical facilities and have greater opportunity to utilize the cardiology care after the new-onset of atrial fibrillation compared to other racial groups [42].

There are multiple potential explanations for the variation in the anticoagulation effect among different racial groups. Different racial groups have different pharmaco-genetics which affects the metabolism of warfarin and dose-response. Studies found that Asian and whites have lower warfarin dose requirements compared to blacks, and this is potentially due to the variation in genotype frequencies among the races [43] [44]. The geographic and cultural differences could be the primary driving source for this kind of variation more than the race itself [45]. Srivastava et al. (2008) have reported that the awareness of their atrial fibrillation medical condition is lower in blacks compared to other races [46]. Multiple factors such as age, gender, socioeconomic status, region, alcohol abuse, pill burden, comorbidities, bipolar disorder and hospitalizations influence 69% difference between whites and blacks [18] [47].

There are two methods in predicting the warfarin doses for different races; these are the race-specific and race-adjusted dosing algorithms. Limdi et al. (2015) demonstrated that the race-specific algorithm is better than the race-adjusted algorithm, and this is because the race-specific algorithm can explain the higher degree of dose variability better than that of the race-adjusted [48]. African American racial group has a lower degree of warfarin dose variability, and both the algorithms can explain this. However, the race-adjusted algorithm is considered further convenient to be used clinically as it can estimate the warfarin dosing for all patients irrespective of their race [49].

Ohara et al. (2019) reported that >50% of Asian patients with VKORC 1-1639G/G or G/A genotypes have sub-therapeutic INR control. Minor allele frequencies of genetic predictors may vary significantly among three racial groups. Asian patients have less INR variability compared to that of the White population, and this could be due to the high prevalence of VKORC1-1639A/A genotypes among this population [50]. CYP2C9*8 is a genotype which is more common among African American patients compared to Asian and White patients, and this may lead to significant INR variability in this specific race [51]. A clinical trial has shown that the mean TTR is lower in genotype-guided dosing compared to clinically guided dosing among black patients, and this indicates that genotype-guided dosing is not always better than the latter [52].

4) Education and other social factors: Chalachew et al. (2019) found that non-clinical factors such as >300 km to medical facility for follow-up, less frequent clinic visits, lower educational level and free warfarin supply source from public institutions were significantly associated with sub-therapeutic INR control. Patients with lower education may have difficulty understanding the nature of the clinical condition, the importance of regular follow-up and INR monitoring [53]. Hospitalizations and outpatient emergency visits may influence INR elevation, and this may be due to adjustment of the warfarin dose with the incidental INR without prior knowledge about the long-term dosing or compliance.

Patients who are dependent on someone’s help for their daily living also have a higher risk of excessive anticoagulation [3] [54]. Patients living in a rural area are very likely to have supra-therapeutic INR control compared to those living in urban areas, and this could be because of lack of accessibility to medical facilities, poverty and social exclusion [35]. In addition, living alone significantly increases the risk of sub-therapeutic INR control by 53% [29]. Individuals who are living alone are very likely to have a poor diet which may affect their vitamin K consumption [55]. This reduction in vitamin K dietary intake increases the risk of poor INR control [56].

5) Smoking and Alcohol: A systemic review has shown that smoking significantly increases the warfarin clearance, which eventually reduces the effect of warfarin [57]. There is no evidence on the interaction between warfarin and nicotine. However, the body absorbs other chemical compounds of the smoke and transports to the liver. The liver produces more enzymes to excrete these toxic chemical compounds. During this process, the liver eliminates more warfarin as well, which eventually increases the warfarin requirements in active smokers. Quitting smoking may also affect the warfarin dosing as these ex-smokers require lower warfarin dose [58]. Therefore, smoking relapses must be questioned to those patients with smoking history plus poor INR control.

Al-Momany et al. (2019) also reported that shisha smoking, as harmful as cigarette smoking, is associated with poor INR control. Smoking shisha is often not consistent, which makes the situation difficult to adjust the warfarin dose and maintains the optimal INR value. In addition, some patients think that smoking shisha is not considered as smoking and prefer not to provide related information to the doctors. Therefore, physicians need to ask the patients specifically on shisha smoking [58].

Alcohol abuse highly increases the risk of over anticoagulation [3] [59]. Excessive alcohol consumption dysregulates liver enzymes production and functions; thus inhibits the warfarin breakdown and also leads to poor compliance to medications [60]. Alcohol abuse also potentially may induce hepatic microsomal enzymes which elevate the warfarin metabolic elimination. As alcohol consumption increases, coagulation factor VII decreases as well [61].

A study of 28,011 patients on warfarin therapy showed that alcohol abuse is significantly associated with poor INR control of iTTR < 60% (Odds Ratio: 2.72, CI: 2.27 - 3.24, p < 0.001) [62]. Excessive alcohol consumption is also a risk factor of bleeding in HAS-BLED score but not included in SAMe-TT₂R₂ and PROSPER scores [63]. ROCKET-AF study showed that alcohol abuse reduced the TTR in patients on warfarin [64]. A critical assessment of the benefits of anticoagulation compared to risk should be conducted before prescribing warfarin. Novel oral anticoagulants (NOAC) could be beneficial for these patients, but there is limited evidence on the efficacy and safety of NOAC in patients with excessive alcohol consumption which suggests further investigation in this issue.

6) Fasting: Fasting can affect the metabolism of warfarin and the volume of warfarin distribution. Studies found a 20% reduction in systemic clearance and central volume of S-warfarin distribution in healthy individuals with 36-hours of fasting. They also identified the area under the concentration curve of S-warfarin, which increased by 20% with fasting at unchanged warfarin therapy. Furthermore, they have shown that the level of the fatty acids in plasma increases after fasting. This increase in plasma fatty acids influences the S-warfarin protein binding [65] [66]. Lai et al. (2014) demonstrated that the risk of supra-therapeutic INR control increases in Muslim patients on warfarin therapy during Ramadan fasting [67].

Katada et al. (2019) also found that patients who were fasting had significantly higher INR value compared to non-fasting patients. The maximum warfarin sensitivity index values were significantly different between these fasting and non-fasting patients. The intra-individual variability of warfarin therapy is significantly affected by fasting [68]. A similar study also described that reduction in dietary intake in critically ill patients increases the hypersensitivity to warfarin. Intravenous infusion of vitamin K₂ had played a vital role to reduce the INR value in these patients [69]. Besides, patients on warfarin therapy with cardio-vascular diseases require more attention during fasting as these patients often have poor nutritional status [70].

7) Weight: Martin-Perez et al. (2019) reported that a lower body mass index (BMI) is associated with an elevation of INR ≥ 5 [35]. A case-control study also found that a BMI < 20 kg/m² increase the risk of supra-therapeutic INR of ≥6. This study identified that a recent loss of a minimum of 2 kg weight is a risk factor of poor INR control [71]. Besides, it has been previously proposed that the clotting factors VII and X reduce with the decrease in weight [72]. Cohen et al. (2019) demonstrated that as the weight increases, INR overshoot decreases which indicates that weight elevation particularly > 90 kg is a protective factor against supra-therapeutic INR [17].

8) Self-management and INR Control: Like weight, patient self-management (PSM) similarly acts as a protective factor as it is an essential strategy in managing the patients with chronic illness including hypertension, diabetes and chronic obstructive pulmonary disease [73]. Multiple PSM programs have been established to assist the management of these diseases. Several randomized control trials have found that PSM is effective and safe to be used to manage a patient [74] [75]. Moreover, PSM is considered as more cost-effective and economical. Hence, channelling more financial resources to educate patients on anticoagulation therapy would be a wise strategy [76]. Self-management via telemedicine programs will be beneficial for well-controlled patients compared to face-to-face appointments in the anticoagulation clinic [77].

PSM significantly reduces the risk of thromboembolism, bleeding and mortality compared to the routine management [78]. Therefore, scientific societies recommended PSM to be used for eligible patients on anticoagulation therapy [79]. PSM correspondingly helps the patients to self-care of their long-term anticoagulation, including INR monitoring and warfarin dose adjustments. The patients can do self-tests and self-adjust the warfarin therapy according to the INR values. Thus, PSM may aid physicians in managing the increasing number of patients on oral anticoagulants [80].

Prochaska et al. (2019) similarly reported that self-management is associated with a reduction in risk of sub-therapeutic INR control by 78%, and this established that self-management acts as a protective factor against poor INR control [29]. This study finding is also consistent with the findings published in the Cochrane database [75]. However, the effectiveness of self-management is questionable in a recurrent sub-therapeutic INR. Grunau et al. (2011) have conducted a study comparing physician management with PSM over a period of 4 months. The mean TTR in physician management and PSM groups is 82% and 80% respectively. During the self-management period, no thromboembolism or major bleedings occur, which shows a promising outcome with PSM [81]. Self-management with active participation in dose adjustments of warfarin will be beneficial for patients with TTR ≥ 75% [82]. Studies have reported that patients, who self-manage, have equal or higher TTR compared to conventional management [83].

Another study has also demonstrated that self-management is associated with lower complications, including thromboembolism compared to conventional management (Odds Ratio: 0.51, 95% CI: 0.31 - 0.85) [84]. In addition, patient self-testing (PST) allows the patients to measure their INR values. However, the warfarin dose-adjustment will be done by physicians [80]. However, Ward et al. (2015) proposed that patients under self-management are more likely to attain better TTR compared to patients under self-testing [85]. Both these PSM and PST can be facilitated by point-of-care testing.

Study Limitations: We did not find any article where the randomized controlled trial has been conducted to analyze how the non-clinical factors affect the INR control in patients receiving warfarin therapy. Hence, we could not conduct any risk of bias assessment to establish transparency of evidence which were synthesized in this review paper.

Though various studies have evaluated the effect of increasing age on the INR control, yet there is no clear unanimous establishment of the association between advanced age and poor INR control or low TTR. That is why careful warfarin dosing and INR monitoring are essential in older patients as they have a higher risk of bleeding. We also found the evidence that female gender is associated with poor INR control, and this gender-specific variation as a non-clinical factor needs to be tackled to ensure their optimal clinical outcomes. The variation in genotype frequencies impacts the warfarin dosing and quality of INR control among different ethnicities. Hence, a race-specific dosing algorithm would be helpful for physicians in prescribing warfarin.

Besides, proper healthcare strategies are required as multiple social factors such as lower education level, living alone, and accessibility to medical facilities may lead to poor INR control. Smoking and alcohol abuse are always being the risk factors of multiple chronic illnesses; poor INR control is not exceptional. Fasting is responsible for increasing the risk of supra-therapeutic INR control, as it affects warfarin metabolism and clearance. Therefore, more attention should be provided to patients on warfarin therapy during any types of fasting, including Ramadan fasting.

A low BMI is associated with supra-therapeutic INR although a high BMI is a protective factor against supra-therapeutic INR, this suggests that the other clinical consequences of excessive weight gain should not be neglected. Like high BMI, PSM acts as a protective factor against poor INR control. Since PSM is more cost-effective, managing eligible patients using self-management would be ideal as PSM is equally or more effective than conventional management.

That is why we recommended that multiple non-clinical factors which influence the INR control in patients on warfarin therapy should not be neglected. Identifying and recognizing these factors before prescribing warfarin would be quintessential to achieve and maintain therapeutic INR. Hence, physicians managing patients on warfarin therapy are encouraged to identify and recognize these non-clinical factors before prescribing warfarin to achieve and maintain the therapeutic INR. Also, a critical assessment of warfarin dosing and the benefits of anticoagulation compared to risk is essential before prescribing warfarin to these patients.

The authors declare no conflicts of interest regarding the publication of this paper.

Ramasamy, T., Pillai, N.K., Yap, C.G. and Jahan, N.K. (2020) Non-Clinical Factors Associated with International Normalized Ratio Control in Patients on Warfarin Therapy: A Review Paper. Open Access Library Journal, 7: e6947. https://doi.org/10.4236/oalib.1106947

AF: Atrial Fibrillation

BMI: Body Mass Index

CI: Confidence Intervals

INR: International Normalized Ratio

IRR: Incidence Rate Ratio

MeSH: Medical Subject Headings

NOAC: Novel Oral Anti-Coagulants

TTR: Time in therapeutic range

PSM: Patient Self-Management

PST: Patient Self-testing

VARIA: Veterans AffaiRs Study to Improve Anticoagulation