Introduction

Atrial fibrillation (AF) is the most common supraventricular arrhythmia. It is estimated that it can involve about 2–4% of the general population, and the prevalence of AF increases with age [1, 2]. Thromboembolic complications, stroke included, are one of the most dangerous implications of AF, and the risk of their occurrence in patients not using anticoagulant treatment is about 5% yearly [3]. Multicenter randomized clinical studies showed the efficacy and safety of non-vitamin K antagonist oral anticoagulants (NOACs) in the prevention of thromboembolic complications in patients with AF. Dabigatran (thrombin inhibitor), apixaban, and rivaroxaban (Xa factor inhibitors), which belong to this group, are characterized by predictable pharmacokinetics enabling their application without monitoring parameters of coagulation and by a lower number of interactions with pharmaceuticals different from vitamin K antagonists. These medications differ from each other in terms of dosage and indications to reduce dosages. Based on the conducted clinical studies, recommendations for choosing the appropriate NOACs dosage in patients with AF were defined. Administering NOACs in a full or reduced dosage is vital for providing patients with efficacious and safe treatment, and the choice of the dosage depends on features such as age, weight, renal impairment, or higher risk of bleeding [4–6]. However, there are some reports that indicate inappropriate administering of NOACs dosages — especially prescription of reduced dosages too frequently, which is against the guidelines. Several studies have shown a higher percentage of patients who received inappropriate reduced NOAC dosages — in Whitworth et al. [7], 33% of patients received a reduced dosage of NOAC against guidelines, and in Barra et al. [8], this percentage was even higher — 46%. An inappropriate dosage reduction of NOAC was associated with reduced effectiveness for stroke prevention without any safety benefits [9].

The study aimed to evaluate the administration frequency of NOAC dosages against the guidelines and analysis of factors predisposing to such a choice in patients with AF included in POL-AF (The Polish Atrial Fibrillation) Registry.

Material and methods

Study group

The Polish Atrial Fibrillation (POL-AF) Registry is a multicenter, prospective, observational study including patients with AF from ten cardiology centers — 7 academic ones, 2 regional hospitals, and one military hospital. The study was registered on clinicaltrials.gov (NCT04419012). The data was gathered from January to December 2019. The aim of the record was to obtain data concerning the clinical characteristics of patients with AF and to evaluate the undertaken steps — especially in terms of thromboembolic prophylaxis. Subsequent patients with AF, hospitalized in the centers for urgent and planned reasons — and who were over 18 years of age and suffered from arrhythmia documented with electrocardiographic examination or medical documents — were added to the record. No clear exclusion criteria were defined to gather a group well-representing Polish cardiological reality; however, patients admitted to the hospital to have ablation due to AF were not included in the record.

Based on the POL-AF record results, patients with AF treated with reduced dosages of NOACs were evaluated in the presented study. Patients receiving full NOACs doses, vitamin K antagonists, and antiplatelet therapy — i.e., those without anticoagulant treatment and with no data about weight or renal function — were excluded from the study (Figure 1).

Figure 1. Flow chart of the study; APT — antiplatelet; NOAC — non-vitamin K antagonist oral anticoagulants; VKA — vitamin K antagonists

Analyzed data

Atrial fibrillation was diagnosed based on medical records or electrocardiographic examination results upon admission to the hospital or during hospitalization. The researchers collected rudimentary data connected with demography, medical record, AF type, laboratory investigation results, and pharmacotherapy. The definitions of comorbidities are presented in Table S1. Creatinine clearance was achieved using the Cockroft-Gault equation. The thromboembolic risk was estimated based on the CHA2DS2-VASc score (congestive heart failure, hypertension, age ≥ 75 years, diabetes, stroke/transient ischemic attack, vascular disorder, age 65–74 years, sex) [10]. The risk of bleeding was defined based on the HAS-BLED score (arterial hypertension, kidney/liver failure, stroke, bleeding, labile international normalized ratio [INR], older age > 65 years, pharmaceuticals/alcohol) [11]. The study was sanctioned by the Bioethical Commission of the Swietokrzyskie Chamber of Physicians in Kielce (104/2018). The commission waived the requirement of obtaining patients’ informed consent.

Appropriateness of NOAC dosage

An evaluation of the appropriateness of reduced NOAC dosage was determined based on guidelines from The European Society of Cardiology, which refer to the treatment of patients with AF, and a summary of product characteristics registered in the European Medicine Agency [1]. The criteria for the approved dosage reduction for each NOAC was as follows: dabigatran 220 mg/day for patients with creatinine clearance (CrCl) 30–50 mL/min; rivaroxaban 15 mg/day for patients with CrCl 15–49 mL/min; apixaban 5 mg/day for patients with more than two of the following: age ≥ 80 years, body weight ≤ 60 kg, or serum creatinine ≥ 1.5 mg/dL. For all NOACs, concomitant use of P-glycoprotein inhibitors was an indication of reduced dosages (Table S2). Inappropriate reduction of the NOAC dosage was defined as fulfilling ≥ 1 criterion of dosage against the guidelines. It referred to patients with a prescribed reduced NOAC dosage, despite qualification for the full dosage (‘underdosed’), and to patients with a prescribed NOAC dosage higher than the one advised or allowed (‘overdosed’). According to the guidelines in force at the time the POL-AF Registry was started, NOAC dosages should be reduced when one or two antiplatelet drugs are used concomitantly; so, we considered that reducing NOAC dosages in patients taking antiplatelet drugs is the correct management.

Statistical analysis

The distribution of quantitative features (age, sex, AF type, medical record, results of laboratory investigation, pharmacotherapy) was verified with the Shapiro–Wilk test. To define the significance of the differences between the groups for particular quantitative features, the Kruskal–Wallis test was used. In the case of features for which the differences appeared to be significant, post hoc Dunn Bonferoni tests were also used. All the qualitative variables were coded in a zero-one system, where 0 means the lack of a particular feature and 1 means its presence. In individual NOAC groups, and later in appropriate/inappropriate dosage groups, stratum weights were calculated and the relationship between variables was estimated using a non-parametric c2 test. To determine the influence of chosen variables on medication dosage, a logistic regression analysis was used (the full model was presented). The effects were shown for the inappropriate dosage, whereas the appropriate dosage was the referential level. The value p ≤ 0.05 was assumed to be statistically significant. The data was analyzed using Statistica 13.3 software.

Results

Baseline characteristics

Among the 1003 patients treated with the reduced dose of NOACs, 409 were prescribed rivaroxaban (40.8%), 275 were prescribed apixaban (27.4%) and 319 were prescribed dabigatran (31.8%). The average age of the patients was 77.9 (± 9.4) years, patients over 74 years accounted for 68.7% of the researched population. Women accounted for 47.9% of the patients. The average result in the CHA2DS2-VASc score was 5.1 points, whereas, in the HAS-BLED score, it was 2.3 points. Renal impairment, defined as creatinine clearance < 60 ml/min, was diagnosed in 640 patients (63.8%). Previous bleeding, including bleeding from the digestive tract, appeared in 64 patients (6.4%) — most frequently in the group of patients treated with apixaban (28 patients, 10.2%), least frequently in patients using rivaroxaban (15 patients, 3.7%). In the studied population, the first-time episode of AF was observed in 61 patients (6.1%). Heart failure (HF) was the most frequent reason for hospital admission; it referred to 268 patients (26.7%) — most of them from the group treated with apixaban (101 patients, 36.7%). Planned reasons for hospitalization, such as electrical cardioversion and ablation of arrhythmia different from AF, were observed most often in the group treated with rivaroxaban — respectively, 67 patients (16.4%) and 21 patients (5.1%). The clinical characteristics of patients included in the study are presented in Table 1.

Assessment of the propriety of using a reduced dose of NOAC

An appropriate NOAC dosage reduction was observed in 761 patients (75.9%), and an inappropriate NOAC dosage reduction was observed in 242 patients (24.1%). An inappropriate reduced dosage was observed in 120 patients (29.3%) treated with rivaroxaban, in 93 patients (33.8%) treated with apixaban, and in 29 patients (9.1%) treated with dabigatran (p < 0.0001). The frequency of an appropriate and inappropriate NOAC dosage reduction in each NOAC group is shown in Table 2, while a comparison of patients treated with appropriate/inappropriate reduced NOAC dosages is shown in Table 3. Patients using inappropriately reduced NOAC dosages, in comparison to patients receiving appropriately reduced dosages, were younger (the average age was 76.9 years vs. 78.3 years, p > 0.05), and the proportion of women in this group was higher (51.2% vs. 46.8%, p > 0.05). The occurrence frequency of HF, vascular diseases, and renal impairment was lower in the group of patients treated with inappropriately reduced dosages than in the group with appropriately reduced dosages (p < 0.05 for all). Moreover, in this group, the number of patients concurrently using antiplatelet treatment was lower compared to the group administered appropriately reduced dosages (0.4% vs. 33.9%, p < 0.05).

The average number of points in the CHA2DS2-VASc score was lower in the group with inappropriately reduced doses (4.8 vs. 5.2, p < 0.05) (Table 3). Figure 2 shows the proportion of patients treated with appropriately/inappropriately reduced dosages according to the CHA2DS2VASc score. Electrical cardioversion and CIED implantation/reimplantation were more frequent reasons for hospitalization among patients treated with inappropriately reduced NOAC doses (p < 0.05 for both).

Figure 2. Prescription pattern of appropriately and inappropriately reduced non-vitamin K antagonist oral anticoagulants (NOAC) dosages based on the CHA2DS2-VASc score

The number of points on the CHA2DS2-VASc score

Percent of patients treated appropriately
and inappropriately reduced NOAC dosages

Inappropriately reduced NOAC dosages

Appropriately reduced NOAC dosages

Predictors of the use of inappropriately reduced dosages of NOACs

The univariate logistic regression analysis found numerous predictors of inappropriately reduced dosages of NOACs prescription (Table S3).

In the multivariable model, factors associated with the selection of the inappropriately reduced dosages of NOACs included the following: age, HF, vascular disease, CrCl < 60 mL/min, and hospitalization due to CIED implantations/reimplantations.

Table 4 demonstrates the predictors of the use of inappropriately reduced doses of NOACs. Independent predictors of using inappropriately reduced doses of NOACs were HF (OR 1.55, CI: 1.08–2.22) and hospitalization due to CIED implantations/reimplantations (OR 2.01, CI: 1.27–3.17). Factors diminishing chances to administer inappropriately reduced doses of NOACs were age (OR 0.98, CI: 0.97–0.998), vascular disease (OR 0.29, CI: 0.21–0.40), and CrCl < 60 mL/min (OR 0.37, CI: 0.27–0.52).

Discussion

The use of NOAC in clinical practice is becoming increasingly common; therefore, numerous international registers are being kept to assess the factors influencing the choice of anticoagulant therapy in the prevention of thromboembolic complications in patients with AF [12–14]. Despite the definite indications to reduce NOACs doses in randomized clinical studies — ARISTOTLE (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) [4], RE-LY (Randomized Evaluation of Long-Term Anticoagulation Therapy) [5] and ROCKET-AF (Rivaroxaban once-daily, direct oral factor Xa inhibition compared with vitamin K antagonism for prevention of stroke and Embolism Trial in Atrial Fibrillation) [6] — the selection of an appropriate NOAC dosage in clinical practice still remains a huge challenge for doctors. Several studies have reported clinical outcomes according to label adherence to NOAC dosage. Using a reduced NOAC dose without any dosage-reduction criteria could lead to a below-par reduction of stroke risk [15]. Steinberg et al. [16] showed that NOAC overdosing was associated with increased all-cause mortality compared to recommended doses, and underdosing was associated with increased cardiovascular hospitalization.

The presented study showed that in 24.1% of patients, inappropriately reduced NOACs dosages were used. Yu et al. [17] showed that 31% of NOAC-treated patients were undertreated. In the population of elderly patients, it was shown that 51% of them received a reduced dosage despite the lack of formal criteria for dosage reduction [18]. The proportion of patients treated with an inappropriately reduced NOAC dose in our study is higher compared to studies of other authors — reduction of a NOAC dose against the guidelines was observed in 14.4% of patients in the study of Ono et al. [19]; in 16.1% in the study of Gustafson et al. [20], whereas in the ORBIT AF II study, an inappropriate dosage reduction was observed in only 9.4% of patients [16]. The presented study is multicenter; it included hospitalized patients, administering a reduced NOAC dose despite the lack of indications defined in the guidelines could result from the presence of different from those commonly acknowledged factors substantively increasing the risk of bleeding (e.g., frailty syndrome, psycho-organic syndrome).

In the discussed study, the highest proportion of inappropriate dosage reduction referred to patients receiving rivaroxaban (29.3%) and apixaban (33.8%), which is reflected in the results of some other authors’ studies. In the study of Ono et al. [19], inappropriately reduced dosages were observed in 12.8% of patients administered rivaroxaban and in 19.6% of patients treated with apixaban, whereas in the study of Gustafson et al. [20], the proportion was even higher — 47.5% and 42.5% for rivaroxaban and apixaban, respectively. One of the probable reasons for using an inappropriate dose reduction of rivaroxaban or apixaban can be an underestimation of renal function caused by different models of CrCl calculation — most laboratories present the CrCl value calculated based on MDRD (Modification of Diet in Renal Disease) formula, whereas customizing NOACs dosages in clinical studies was based on CrCl result achieved from the Cockroft–Gault equation [21, 22].

In the presented study, CrCl < 60 mL/min was shown to be a significant factor in diminishing the risk of using an inappropriate NOAC dose. It has been proven that using NOAC was associated with a reduced risk of thromboembolic and hemorrhagic complications compared to warfarin in patients with mild and moderate renal impairment [23–25]. Recent studies have also confirmed the effectiveness of reduced doses of NOAC in patients with concomitant chronic kidney disease, showing a promising benefit-risk balance in populations at high risk for cardiovascular complications [26].

In the POL-AF registry, the chance of using the inappropriately reduced NOAC dosage decreases by one percent every year. Older age was an independent factor in diminishing the risk of prescribing inappropriately reduced doses of NOACs in our study. Such conclusions were also drawn in the study of Jacobs et al. [27]; however, for the age group of ≥ 80 years — age was also a predictor of administering the appropriate reduction of doses for all researched NOACs. Interestingly enough, in the presented study, renal impairment and vascular disorder also appeared to be factors that diminished the risk of inappropriate dosage reduction. The study of Ono et al. [19] demonstrates that both older age (> 65 years), as well as renal impairment (CrCl < 60 mL/min), was independently connected with the risk of using inappropriately reduced doses — while vascular disease was a predisposing factor for prescribing an appropriately reduced dose of NOAC. The reason for the appropriate NOAC dosage reduction in patients with vascular disorders in our study could be the simultaneous use of antiplatelet drugs. In the methodology, we assumed that reducing NOAC doses in patients taking antiplatelet drugs is the correct management — according to the guidelines in force at the time the POL-AF Registry was started, which recommend considering a NOAC dose reduction when one or two antiplatelet drugs are used concomitantly. However, guidelines from the second half of 2019 allow using full doses of NOACs with simultaneous antiplatelet therapy.

There are many potential reasons why clinicians might prescribe inappropriate NOAC doses. In the discussed study, two factors predispose to prescribing inappropriately reduced dosages of NOACs. The first of them is hospitalization due to CIED implantation/reimplantation. The probable reason for such a study result could be doctors’ fear of a hematoma when administering full doses of NOACs directly after the surgery. Because the influence of hospitalization reasons on the potential of using improper decreased doses was not examined in the papers that are readily available, it is not possible to compare the acquired result with other authors’ investigations.

Another factor predisposing to an inappropriate prescription of reduced dosages of NOACs in our study is HF, as in the study by Ono et al. [19]. The reason for the inappropriate reduction of NOAC dosage in patients with HF could be the fear of worsening renal function in patients concomitantly taking diuretics in the absence of proper fluid balance control. However, it should be remembered that HF is one of the factors in the CHA2DS2-VASc score, which measures the risk of thromboembolic complications in patients with AF and, therefore, the selection of the appropriate NOAC dosage is important to ensure effective treatment of patients.

What is important in the presented study is the risk of bleeding estimated based on the HAS-BLED score was not a significant predictive factor of using inappropriately reduced NOAC doses, just as in the studies of Ono et al. [19] and Jacobs et al. [27]. In the SAFE-NOACS study, past hemorrhagic complication was also not a factor predisposing to an inappropriate NOAC dose correction [28]. The risk of bleeding during anticoagulant treatment is higher in some of the patients; however, it should not be the reason for an inappropriate reduction of NOAC dosages because it can increase the risk of thromboembolic complications in patients.

Conclusions

In the group of hospitalized patients with AF treated with a reduced dosage of NOAC, 24.1% of them inappropriately reduced the dosage of NOAC prescription, most frequently in patients treated with apixaban and rivaroxaban. The factors predisposing to the prescription of an inappropriate reduced dosage of NOAC were HF and hospitalization due to CIED implantation/reimplantation. Age, vascular disorder and renal impairment were independent predictors lowering the risk of prescribing an inappropriate reduced dosage of NOAC.

Acknowledgments

The POL-AF registry was initiated on the “Club 30” Scientific Platform of the Polish Cardiac Society. Investigators other than those listed as Authors include: Anna Michalska (Kielce), Paweł Krzesiński (Warszawa), Katarzyna Karoń (Warszawa), Monika Szewczak (Warszawa), Wiktor Wójcik (Warszawa), Michał Niedźwiedź (Warszawa), Bartosz Krzemiński (Grodzisk Mazowiecki), Arkadiusz Sokołowski (Grodzisk Mazowiecki).

Conflict of interest

IGG — speaker for Bayer and Boehringer-Ingelheim; AKC — speaker for Bayer; JBednarski, AM and BWK — speaker for Bayer, Boehringer-Ingelheim and Pfizer; ATK — speaker for Boehringer-Ingelheim; MWełnicki — speaker for Bayer and Pfizer; other authors have no conflict of interest to declare.

Data availability

The data used to support the findings of this study is available from the corresponding author upon request.

Funding statement

Project financed under the programmed of the Minister of Science and Higher Education, called ‘Regional Initiative of Excellence’, in the years 2019–2022, Project no. 024/RID/2018/19, amount of financing 11,999,000 PLN.

→