The Influence of High Body Mass Index (BMI > 35 kg/m²) on Apixaban Plasma Concentration in Patients with Atrial Fibrillation

Apixaban, one of four direct oral anticoagulants (DOACs), offers an option for managing and preventing thromboembolic events in patients with atrial fibrillation. Apixaban, a direct factor Xa (FXA) inhibitor, is administered in a fixed-dose regimen for most patients, limiting the need for routine blood level monitoring. Dosage adjustment, when required, is based on renal function, age, and body weight less than 60 kg [1‐6].

Obesity is associated with physiological changes, such as an increase in body mass including adipose tissue, lean body tissue, major organs’ mass, and muscle mass, and in body fluids [7‐9]. Furthermore, changes have been observed in plasma proteins, metabolizing enzymes, and blood flow to organs [7, 8, 10, 11]. Notably, the kidneys undergo both structural and functional adaptations, with an augmented effective renal plasma flow, increased albumin secretion, and an approximately 60% rise in glomerular filtration rate (GFR) [12, 13]. Liver function may be influenced by alterations in blood flow and the accumulation of fat within the liver [7]. Moreover, these changes extend to the binding of drugs to plasma proteins, which is affected by shifts in the level of alpha 1-acid glycoprotein and albumin [9, 14, 15].

Obesity-related physiological changes have the potential to influence the pharmacokinetics of various drugs [7, 8, 10, 13, 16‐18]. Nonetheless, there are insufficient data on the pharmacokinetics of apixaban in the obese population, and the existing studies evaluating its efficacy and safety in this specific group remain relatively limited. Recent studies, such as those by Chin-Hon et al., have extensively analyzed the safety and efficacy of DOACs in patients with extreme body weights [29]. Notably, the pivotal ARISTOTLE trial, which demonstrated apixaban’s efficacy in patients with atrial fibrillation, reported a median weight of 82 kg, with an interquartile range spanning from 70 to 96 kg. While the study stratified data based on body weight, it did not recommend dose adjustments based on weight alone. Dose adjustments were primarily guided by renal function and specific patient criteria, such as age and weight less than 60 kg, as noted in the revised prescribing information labeling (March 2014), as detailed in Sect. 2.7 of the labeling [19].

A review of the literature on the effect of obesity on apixaban treatment identified a limited number of small-scale studies that explored both the pharmacokinetic and pharmacodynamic aspects, encompassing efficacy and safety [20‐30].

Recent studies, such as those by Chin-Hon et al., Rueda-Camino et al., and Sperry et al., have extensively analyzed the safety and efficacy of DOACs in patients with extreme body weight [29‐31]. The researchers noted that further studies are needed, including a recommendation by Chin-Hon et al. for monitoring drug levels in obese patients.

Due to the limited availability of robust clinical data, the International Society on Thrombosis and Haemostasis (ISTH) in 2016 [32] and Australian hematology experts [2] recommended against using DOACs in patients with a body mass index (BMI) > 40 kg/m² or high body weight weighing exceeding 120 kg.

Nevertheless, in 2021, subsequent to a series of studies aimed at examining the efficacy and safety of apixaban in individuals with extreme BMIs [33, 34], the European Heart Rhythm Association practical guide on the usage of non-vitamin K antagonist oral anticoagulants in patients with atrial fibrillation recommended the cautious use of DOACs, including apixaban, considering plasma level measurement for patients with these specific BMI and body weight profiles [35]. In addition, the American Heart Association, American College of Cardiology, and Heart Rhythm Society recommended the monitoring of plasma levels of DOACs in patients with extreme obesity, defined as BMI > 35 kg/m² or body weight > 120 kg [36].

The primary aim of this study was to investigate the potential impact of BMI ≥ 35 kg/m² on apixaban trough plasma concentrations (C_min) in patients with atrial fibrillation, focusing on BMI as a distinct factor influencing pharmacokinetics, which may offer insights beyond those provided by studies examining broader weight categories.

This prospective cohort study enrolled patients with atrial fibrillation treated with appropriate doses of apixaban. We enrolled 30 patients with a BMI ≥ 35 kg/m² and 30 patients with a BMI < 35 kg/m² hospitalized in the internal medicine departments of HaEmek Medical Center or treated as outpatients. The study was performed between May 2019 and January 2020. Apixaban C_min at steady state were compared between patients with a BMI ≥ 35 kg/m² and patients with a BMI < 35 kg/m².

Patients were considered at steady state if they met one of the following criteria: either they had received a minimum of four consecutive doses of apixaban during their hospitalization, and before their recruitment to the study, or they had initiated treatment before hospitalization and their medication possession ratio (MPR) was 80% or higher over the 6 months leading up to their hospitalization, or since starting apixaban therapy, whichever was of shorter duration. MPR serves as an indirect measure to determine the compliance of patients over a defined period [37]. We assumed that achieving a high MPR value (≥ 80%) corresponds to a drug plasma concentration that reflects consistent medication adherence leading up to hospitalization and mirroring the steady-state condition of patients who had received four consecutive doses.

The study was approved by the institutional review board (Helsinki Committee) of the HaEmek Medical Center.

The sample size of 60 patients in total was empirically determined due to a lack of available data for this comparison.

The dosages of apixaban that were prescribed to all patients who were included in our sample fitted the current clinical recommendations and prescribing instructions for apixaban [43].

We excluded patients with vomiting and diarrhea, anasarca, chronic kidney disease grade V (estimated GFR [eGFR] calculated with the Chronic Kidney Disease Epidemiology Collaboration creatinine equation [eGFR-EPI] less than 15 ml/min/1.73 m²), hepatic failure, or bariatric surgery.

Weight and height were measured for all patients, and their BMI was calculated by dividing their weight (kg) by the square of height (m²) [38].

Apixaban C_min was measured before the morning apixaban dose (12 h after the last dose). Apixaban sampling time for C_min measurement adhered to the recommendations outlined by the European Society of Cardiology (ESC) for patients with obesity [6]. Following collection, blood samples were promptly transported to the hematological laboratory at the hospital for the quantification of apixaban plasma concentration measurement, and the tests were conducted within a few hours of sample collection. HemosIL Liquid anti-Xa^® was the commercial assay employed by the lab for this purpose [39].

Additional data were also collected from medical records, encompassing patients’ demographics, comorbidities, and kidney function (assessed using the eGFR-EPI) [40]. Chronic medications and those administered during the hospitalization period were reviewed to rule out potential interactions that might influence apixaban serum level.

Furthermore, the CHA₂DS₂-VASc score was calculated using demographic and morbidity data to predict thromboembolic events risk in patients with non-valvular atrial fibrillation [41].

Between May 2019 and January 2020, a total of 60 patients were recruited to the study, with 58 patients being hospitalized in the internal medicine departments of HaEmek Medical Center and two outpatients. One patient was excluded from the analysis due to a significant delay of approximately 17 h between the last apixaban dose and the collection of their blood sample. Consequently, the final cohort of this analysis consisted of 59 patients. Among them, 29 patients with a BMI ≥ 35 kg/m² were categorized as the “obesity group”, while the remaining 30 patients with a BMI < 35 kg/m² formed the “control group.”

The majority of patients were prescribed 5 mg of apixaban twice daily—specifically, 16 out of 30 patients (53.3%) in the control group and 24 out of 29 patients (82.8%) in the obesity group (Table 1).

The mean body weight of patients in the control group was 75.1 ± 13.2 kg (range spanning from 46.8 to 96.6 kg). In contrast, the obesity group had a substantially higher mean body weight of 106.1 ± 17.4 kg (ranging from 72.2 to 147.4 kg) (P < 0.001). The median BMI of the control group stood at 29.4 kg/m² (ranging from 19.6 to 34.9 kg/m²), while that of the obesity group was notably higher, at 39.1 kg/m² (ranging from 35.1 to 56 kg/m²) (P < 0.001).

Among the observed comorbidities, diabetes was more prevalent in the obesity group, and no other substantial disparities in comorbidity rates were found. Furthermore, there was no statistically significant difference in the CHA₂DS₂-VASc score between the two groups (P = 0.981).

All medications taken for either chronic conditions or acute illnesses were reviewed; none of the patients had received a strong inhibitor of P-glycoprotein (P-gp) or cytochrome P450 (CYP) 3A4, nor a CYP3A4 inducer. Notably, five out of the 29 patients in the obesity group had received a drug classified as a moderate inhibitor of P-gp and CYP3A4, with four of them treated with amiodarone and one with verapamil. In comparison, four out of the 30 patients in the control group were treated with amiodarone (Table 1).

No statistically significant difference was observed in the median C_min among patients receiving 2.5 mg of apixaban twice daily between the obesity group and the control group (106 ng/ml and 123.2 ng/ml, respectively) (P = 0.18). Conversely, for patients receiving 5 mg twice daily, the median C_min was notably lower in the obesity group compared to the control group (148.9 ng/ml vs. 209.1 ng/ml, respectively) (P = 0.044), reflecting a 29% reduction in the obesity group. Figure 1 illustrates the distribution of apixaban C_min by dose and BMI, highlighting the differences between these groups. These trends are depicted in Fig. 2, showing the relationship between obesity, BMI, and C_min at both dose levels. This analysis confirmed a trend of decreasing median C_min as BMI or body weight increases (Table 2).

The median apixaban C_min of patients included in this study were compared to results from an apixaban phase 3 clinical trial [43] (see Table 2).

The median C_min in patients receiving 5 mg twice daily, whether in the control group or the obesity group, were found to be higher than the manufacturer’s values. However, all those values remained within the predicted range.

Specifically, the C_min exceeded the manufacturer’s predicted 95th percentile in 44% of patients in the control group and 16% of those in the obesity group administered 5 mg twice daily. However, the C_min exceeded the 95th percentile in 21.5% of patients administered 2.5 mg twice daily, all in the control group.

Various independent variables among the study participants were examined to determine their potential influence on the C_min of apixaban, such as age, sex, weight, BMI, and kidney function. Table 3 summarizes the results according to each dosing regimen, highlighting the correlation of each individual variable with C_min.

Apixaban C_min were not correlated with the variables tested in patients administered 2.5 mg twice daily.

Apixaban C_min directly correlated with patient age in patients administered 5 mg twice daily and inversely correlated with patient weight and GFR (see Table 3). Interestingly, no correlation between BMI and C_min was found.

Patients with extraordinarily high C_min were examined for possible explanations.

Seven of the 16 patients (44%) in the control group who were on 5 mg twice daily exceeded the 95th percentile of C_min established during the apixaban phase 3 trial, which was set at > 230 ng/ml. Their mean C_min was 296.7 ng/ml.

Four of the 24 patients (16.7%) in the obesity group who were on 5 mg twice daily exceeded the 95th percentile of C_min in the apixaban phase 3 trial (> 230 ng/ml). Their mean C_min was 307.7 ng/ml.

Among the total of five patients who exhibited C_min values exceeding 300 ng/ml, all were prescribed 5 mg apixaban twice daily. A possible explanation for these high values was found in four cases: the first patient had one borderline criterion for dose reduction (borderline weight of 64 kg) and he was 90 years old; the second patient had borderline eGFR values (37 ml/min/1.73 m²); the third patient had low body weight (54 kg) (dose reduction does not apply in patients fulfilling only one criterion according to the guidelines); and the fourth patient had been administered a loading dose of 800 mg amiodarone. The fifth patient had no obvious reason for his extraordinarily high C_min.

Of the 14 patients receiving 2.5 mg twice daily in the control group, three (21.4%) exceeded the expected 95th percentile of C_min (> 162 ng/ml). Intriguingly, no obvious explanation could be identified. None of the patients in the obesity group on the same dosing regimen (encompassing a total of five patients) exceeded the range established in the phase 3 apixaban trial.

In our study, there were five patients with body weight exceeding 120 kg (129, 133, 133.1, 135, 147.4 kg) (Table 4). Among these patients, four were on 5 mg twice daily; their mean C_min was lower, measured at 122.6 ng/ml, when compared to the mean C_min of patients in the obesity group who were on 5 mg twice daily, measured at 165.2 ± 83.3 ng/ml. This difference amounted to a 44% reduction compared to the control group receiving the same 5-mg twice-daily dosage. However, all maintained C_min within the predicted values. Their mean age was 66.5 years, mean eGFR was 66 ml/min/1.73 m², and all were males except for one female.

Overall, 14 patients had a BMI > 40 kg/m²; of them, 12 were on 5 mg twice daily, and the remaining were on 2.5 mg twice daily. The mean C_min observed within this subset was 159.9 ± 64.9 ng/ml, all within the predicted range. Notably, the mean BMI was 44.2 ± 4.4 kg/m².

To neutralize the effect of drug–drug interactions, with amiodarone or verapamil potentially affecting apixaban C_min, mean C_min was calculated excluding patients who received these drugs. The mean C_min of patients in the obesity group who received 5 mg twice daily was 32% lower than that of patients in the control group who received the same dose of apixaban (150.5 ng/ml compared to 219 ng/ml, respectively).

Our study aimed to establish whether BMI ≥ 35 kg/m² exerts an influence on the C_min of apixaban in patients with atrial fibrillation. While Granger et al. and Zhao et al. suggest that standard dosing regimens are effective across various BMI ranges, they do not specifically recommend dose adjustments based on body weight or BMI. The rationale for this study stems from the observed physiological changes in obese individuals that might impact drug pharmacokinetics and pharmacodynamics [19, 44]. Obesity causes an increase in the volume of distribution and the volume of blood and an increase in cardiac function and lean body weight [6‐8, 17]. Hepatic flow is changed and the chronic inflammation and cytokines released influence some CYP expression [10, 45]. This may lead to a decrease in the activity of CYP3A enzymes [7, 16, 18], while other enzymes such as CYP2D6, CYP2C19, CYP2C9, and CYP1A2 tend to display increased activity [32]. Twenty-five percent of orally administered apixaban is metabolized by the liver, mainly by CYP3A4/5, with a little involvement of CYP1A2, 2C8, 2C9, 2C19, and 2J2. Therefore, the effect on the liver metabolism of apixaban cannot be predicted. The prior studies mentioned above do not specifically address dose adjustments based on body weight or BMI, and there is an opportunity to explore whether such adjustments might be beneficial in certain sub-populations.

Apixaban concentrations have been measured in a limited number of studies. In one study, participants with a body weight over 120 kg (involving 19 participants) displayed an approximately 31% reduction in apixaban’s maximum plasma concentration (C_max) compared to the controls (comprising 16 participants) who weighed 65–85 kg [21]. In a prospective study conducted from 2017 through 2019, apixaban plasma concentrations were measured in patients weighing > 120 kg and administered 5 mg twice daily. Peak and trough concentrations were measured in 14 and 11 patients, respectively [20]. None of these concentrations fell below the 5th percentile range provided by the manufacturer. Furthermore, there was no direct linear correlation between drug concentration and either body weight or BMI.

In a post hoc analysis of the ARISTOTLE trial, encompassing 18,193 patients, results were stratified by body weight, and 982 patients (5.4%) fell into the high weight category (> 120 kg) [46]. Remarkably, the researchers concluded that apixaban, compared to warfarin, proved to be both effective and safe for patients with extreme obesity, although the percentage of participants weighing over 140 kg was extremely low [47].

The study by Cin-Hon et al. (2023) evaluated the efficacy and safety of treatment with DOACs compared to warfarin for the indications of atrial fibrillation and venous thromboembolism (VTE) in patients with extreme body weights. The study included 492 patients, with 101 receiving apixaban, among whom ten had a BMI over 40. The results demonstrated that apixaban had a lower rate of VTE and bleeding compared to warfarin, particularly in obese patients, suggesting its efficacy and safety in this population [29].

The study has limitations regarding the sample size within the subgroups, and medication adherence outside the hospital is unknown, with no reference to therapeutic changes during the follow-up period. The researchers noted that further studies are needed, including recommendations for monitoring drug levels in obese patients.

Al-Aieshy et al. conducted a similar study, examining levels for apixaban 5 mg administered for diagnoses of atrial fibrillation, atrial flutter, and VTE in 40 patients with an average BMI of 39.4 compared to 40 patients with an average BMI of 23.4. Trough and peak levels were measured after at least 3 consecutive days of apixaban administration. C_min were higher in the obese group compared to the control group [28]. Unlike the study by Al-Aieshy et al., our results showed lower C_min in the obese group receiving apixaban 5 mg for atrial fibrillation compared to the control group. This study presents findings opposite to ours in terms of C_min and differs from previous studies (cited in our work), showing significant variability among patients, similar to our findings.

This is one of a very few prospective studies examining blood levels of apixaban in patients with atrial fibrillation compared to a control group to study the influence of obesity and BMI on C_min of the drug. The results of our study are in alignment with the findings of a recently published large database study in 2023, stating that DOAC treatment is effective and safe for stroke prevention in atrial fibrillation patients across the BMI spectrum, including extreme obesity [48]. Prior studies that attempted to evaluate the effect of obesity on plasma concentration of apixaban were retrospective and confined by small sample size. A prospective study cited by the clinical guidelines of the ISTH and drug brochures measured apixaban plasma concentration after administration of single oral dose of 10 mg apixaban in 18 healthy individuals [21]. Another prospective study was published in 2020 investigating patients receiving apixaban at full and appropriate doses and in steady state. However, the study lacked a comparable control group with similar characteristics [20], and the prospective study of Al-Aieshy et al., which was performed after our study (between January 2020 and November 2022), presented opposite results to ours.

Our study results indicate a 29% reduction in median C_min among patients administered 5 mg twice daily with a BMI ≥ 35 kg/m². These findings are consistent with a previous study in which the C_max values of apixaban in individuals with high body weight were 31% lower than in the control group, whose weights ranged from 65 to 85 kg [21].

In the current study, we identified additional variables that affect C_min, including weight, age, and GFR. However, no correlation was found between BMI and apixaban C_min in the multivariate analysis. BMI in patients with high body weight does not seem to be an appropriate indicator for predicting apixaban levels, whereas the patient’s weight emerges as a more appropriate indicator.

Despite the decrease in C_min among patients with high body weight, the majority of the results were within the predicted range, and it is unlikely that a dose adjustment is warranted for patients with a BMI ≥ 35 kg/m². Nonetheless, further prospective studies are needed to validate the results.

While the median values in all subgroups were within the predicted range of the drug, their numerical values were higher in both study groups. Patients on 5 mg twice daily with a BMI < 35 kg/m² had a median concentration of apixaban that was twice as high as the manufacturer's predicted value, and in patients with a BMI ≥ 35 kg/m², it was 1.5 times higher. These higher values may be attributed to the different commercial kit used in our hospital. Specifically, we used the HemosIL Liquid anti-Xa kit, whereas apixaban concentrations cited in several databases were assessed using the Rotachrom^® Heparin Anti-Xa assay. While prescribing information acknowledges that different kits can yield different results [49], there are no existing data on differences in apixaban plasma concentration results when using the two different kits.

Our patients were older than the patients in the ARISTOTLE cohort (median age of 70 years) but follow the same age range. Previous studies have indicated a direct relationship between drug concentration and patient age, with a 30% increase in the area under the curve (AUC) observed in patients over 65 years old compared to those patients under 65 years [26, 50]. Our study similarly revealed a substantial direct correlation between C_min and patient age. In addition, our patients displayed poorer renal function in comparison to the ARISTOTLE cohort, where only 15% of participants in the ARISTOTLE trial had chronic kidney disease at stage 3. Since 27% of the apixaban dose is eliminated through the kidneys, impaired renal function may lead to increased plasma concentrations [51, 52]. In our study, within the 5-mg twice-daily dosage group, an inverse relationship between plasma concentration and eGFR was observed.

Patients receiving 5 mg twice daily with C_min exceeding the optimal threshold (concentration higher than 300 ng/ml) represent “real-world” patients in whom dose reduction is not required according to the criteria investigated in the large phase 3 trials. However, surpassing the effective therapeutic range concentrations carries the potential for an increased risk of bleeding, which could impact the treatment’s safety. Therefore, prospective studies are needed to evaluate the necessity of plasma level measurement in such clinical situations.

The results presented here indicate that in a real-life patient setting, especially among complex hospitalized patients with various factors contributing to increased drug plasma concentrations, different values may be obtained than the ones shown in patients selected for clinical trials.

We did find a 44% difference in mean C_min between patients weighing over 120 kg and those with a BMI < 35 kg/m². However, the number of patients weighing over 120 kg was very small (five patients); therefore, no conclusion can be drawn regarding their cases.

Our study has limitations. Firstly, the statistical analysis was performed on 59 patients of the 60 initially recruited, although we assume that this had minimal effect on our results. Secondly, our inclusion of patients receiving either 5 mg or 2.5 mg apixaban twice daily resulted in a small sample size in each subgroup, especially in the 2.5-mg twice-daily group. Consequently, it is difficult to derive conclusive findings, due to the limited sample size. To validate these results, further studies on larger sample sizes are necessary.

In conclusion, our study showed no significant correlation between BMI and apixaban C_min in patients with atrial fibrillation. Nevertheless, a noteworthy 29% lower median C_min of apixaban was evident in patients with a BMI ≥ 35 kg/m² compared to those with a BMI < 35 kg/m². Importantly, all C_min remained within the predicted range for unselected patients. While our findings suggest that BMI may not serve as an appropriate indicator for predicting apixaban levels, further prospective studies specifically focusing on patients with extreme obesity, particularly those with a body weight exceeding 120 kg, are warranted to assess the clinical significance of this observed difference in terms of efficacy and safety.